Method for evaluating durability of arch rib of concrete-filled steel tube arch bridge by considering toughness

Abstract

According to the durability evaluation method for the arch rib of the concrete-filled steel tube arch bridge considering the toughness, comprehensive evaluation on the durability of the arch rib is realized by combining a sensor network, nondestructive testing, a toughness evaluation test, environmental monitoring and an artificial intelligence analysis technology. A vibrating wire strain gauge sensor is arranged on an arch rib to collect strain data in real time; detecting internal defects and temperature abnormity by utilizing ultrasonic detection and infrared thermal imaging technologies; performing an impact test through an impact testing machine, and evaluating the toughness of the arch rib; a weather station is installed to monitor environmental parameters such as temperature, humidity and wind speed, and external factors influencing the performance of the arch rib are obtained. According to the method, data are subjected to MATLAB cleaning processing, a multi-layer neural network and an LSTM time sequence model are constructed through TensorFlow, arch rib durability scores and residual life prediction are generated, and high-risk area maintenance suggestions are provided for evaluation results. According to the method, multi-source data are comprehensively integrated, and accurate evaluation and long-term safety maintenance of the health state of the arch rib are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of durability assessment and engineering maintenance of bridge structures in the field of civil engineering, and particularly relates to a method for assessing the durability of the arch rib of a concrete-filled steel tube arch bridge considering toughness. Background Art

[0002] Concrete-filled steel tube arch bridges are widely used in bridge engineering due to their excellent load-bearing capacity and durability, especially having significant advantages in long-span bridge structures. However, as the main load-bearing member of the structure, the long-term durability of the arch rib of a concrete-filled steel tube arch bridge directly determines the safety and service life of the bridge. Therefore, the accurate assessment of the durability of the arch rib and the monitoring of its health status have become important research directions.

[0003] Currently, the existing durability assessment methods mainly focus on the following aspects:

[0004] (1) Static loading test

[0005] By conducting static loading tests on bridge structures, their load-bearing capacity and deformation conditions are evaluated. This method usually combines finite element analysis to simulate the behavior of the bridge under various load conditions. Although the static loading test can evaluate the overall load-bearing capacity of the bridge structure, the following problems need to be solved urgently:

[0006] Limitations: The applicability of the static loading test is limited, and it can only reflect the short-term performance of the structure, unable to conduct dynamic assessment of long-term durability.

[0007] Lack of toughness assessment: This method mainly focuses on the strength and stiffness of the structure, ignoring the important role of toughness in the seismic and impact resistance of the structure.

[0008] (2) Non-destructive testing techniques

[0009] Non-destructive testing techniques (such as ultrasonic testing, infrared thermography) can evaluate the health status of the arch rib by detecting internal defects and surface damage without damaging the structure. Common methods include:

[0010] Ultrasonic testing: Used to identify defects such as cracks and cavities inside the arch rib.

[0011] Infrared thermography: By detecting abnormal surface temperature distributions, possible crack or corrosion areas are identified.

[0012] Existing problems:

[0013] Insufficient assessment accuracy: Existing non-destructive testing methods rely on a single signal source for defect type judgment and fail to conduct comprehensive analysis by combining multi-source data, resulting in possible deviations in defect identification results.

[0014] Lack of integrated analysis: The integration of ultrasonic testing and infrared imaging results is insufficient, making it difficult to form highly consistent evaluation results.

[0015] (3) Fatigue test

[0016] By conducting repeated loading tests on the arch rib material or small models, simulating the fatigue behavior of the bridge under long-term service conditions, and evaluating its durability. This method reveals the fatigue performance of the material through laboratory data. However:

[0017] Long experimental cycle: Fatigue tests require long-term loading, with a long cycle and high cost, making it difficult to achieve large-scale engineering practical applications.

[0018] Limited to the material level: Fatigue test data is mostly used for the performance evaluation of the material itself and is difficult to directly reflect the durability of the overall structure.

[0019] (4) Long-term monitoring system

[0020] By embedding sensors in the bridge structure to monitor parameters such as stress, strain, temperature, and humidity in real time during the bridge's use to obtain the health status of the structure. This method realizes the monitoring of the bridge's long-term service status, but has the following deficiencies:

[0021] Single monitoring index: Only single parameters such as strain, temperature, humidity, etc. are monitored, and it fails to combine with material properties, defect evaluation, and environmental conditions.

[0022] Insufficient data analysis: Existing methods mostly rely on manual analysis or simple algorithms and are difficult to efficiently process large amounts of data, resulting in insufficient real-time performance and evaluation accuracy.

[0023] In summary, the existing durability evaluation methods have the following technical problems:

[0024] (1) Lack of toughness evaluation

[0025] Current evaluation methods pay more attention to the strength and stiffness of the arch rib, ignoring the key index of toughness. Toughness plays an important role in dealing with extreme working conditions such as earthquakes and impacts and is an important parameter for the safety performance of the bridge structure.

[0026] (2) One-sidedness of evaluation results

[0027] Existing technologies often rely solely on a certain detection method (such as non-destructive testing or static loading testing) and fail to effectively integrate multiple detection methods and data sources, resulting in possible one-sidedness of evaluation results and difficulty in comprehensively reflecting the actual durability of the arch rib.

[0028] (3) Insufficient data processing ability

[0029] With the wide application of sensor networks and monitoring devices, a large amount of data has been generated. Traditional manual analysis and simple data processing methods can no longer meet the needs of bridge health monitoring. Especially in the case of large amounts of data, the data processing efficiency is low and the analysis accuracy is insufficient.

[0030] (4) Lack of comprehensive consideration of environmental factors

[0031] Environmental factors (such as temperature, humidity, wind speed, rainfall, etc.) have a significant impact on the durability of bridges. However, existing evaluation methods fail to incorporate environmental parameters into the overall evaluation system, resulting in insufficient comprehensiveness of the evaluation results.

[0032] (5) Maintenance suggestions are not specific enough

[0033] The results of existing evaluation methods usually stay at the macroscopic level, only putting forward overall suggestions, lacking targeted and operable maintenance measures, and unable to effectively guide the actual maintenance work of bridges. Summary of the Invention

[0034] In view of the problems existing in the above-mentioned prior art, the present invention provides a method for evaluating the durability of the arch rib of a concrete-filled steel tube arch bridge considering toughness. The purpose is to achieve a comprehensive evaluation of the strength, stiffness and toughness of the arch rib by integrating sensor network, non-destructive testing, toughness evaluation and environmental monitoring data, and comprehensively improve the scientificity of the evaluation system.

[0035] To achieve the above purpose, the specific scheme of the present invention is as follows:

[0036] A method for evaluating the durability of the arch rib of a concrete-filled steel tube arch bridge considering toughness, comprising the following steps:

[0037] S1, Sensor network layout: Install strain gauge sensors at key parts of the arch rib of the concrete-filled steel tube arch bridge, and connect the strain gauge sensors to the data acquisition system through shielded cables to collect strain data in real time through the data acquisition system and calibrate regularly;

[0038] S2, Non-destructive testing: Use ultrasonic detectors and infrared thermal imagers to conduct non-destructive testing and marking on internal defects and surface temperature anomalies of the arch rib;

[0039] S3, Toughness evaluation: Extract samples from non-critical stress-bearing areas of the arch rib, conduct impact tests on the samples using an impact testing machine, record the impact energy absorption and toughness experimental data of the failure mode of the samples, and establish a comprehensive toughness scoring model for the arch rib in combination with the strain data and non-destructive testing results in step 1;

[0040] S4, Environmental parameter monitoring: Install weather stations around the bridge and obtain environmental monitoring data such as temperature, humidity, wind speed, and rainfall. Use time series analysis methods to perform correlation analysis on the environmental monitoring data, strain data, and non-destructive testing results to generate time series diagrams and correlation analysis diagrams;

[0041] S5, Data integration and durability scoring: Integrate vibrating wire strain gauge sensors, non-destructive testing, toughness assessment, and environmental monitoring data, and perform multi-dimensional data cleaning, normalization, and modeling to generate a durability assessment report;

[0042] S6, Remaining life prediction: Based on an artificial intelligence prediction model, comprehensively score the durability score and toughness status of the arch rib, and generate a remaining life prediction report in combination with the trend of the comprehensive score;

[0043] S7, Feedback and maintenance suggestions: Identify high-risk areas based on the durability assessment report and provide suggestions for long-term reinforcement, monitoring, and maintenance of the bridge.

[0044] Furthermore, the key parts of the arch rib of the concrete-filled steel tube arch bridge described in step 1 are the mid-span position, arch feet, and transverse connection nodes of the concrete-filled steel tube arch rib; the strain gauge sensor is a vibrating wire strain gauge sensor; the strain data is recorded at a frequency of once per hour; the time for regular calibration is to perform initial reference calibration every 24 hours.

[0045] Furthermore, the ultrasonic detector scans the surface of the arch rib at a frequency of 2 MHz to identify internal cracks and cavities; the infrared thermal imager performs temperature imaging on the surface of the arch rib, records the temperature distribution, and marks the temperature anomaly areas.

[0046] Furthermore, the sample is processed into an impact specimen of 10 mm × 10 mm × 55 mm; the impact speed of the impact testing machine is set to 5 m / s, and the impact energy is 300 J; the failure modes are brittle fracture and plastic deformation.

[0047] Furthermore, the steps for establishing the comprehensive scoring model of the arch rib in step S3 are as follows:

[0048] S31, Data normalization processing: Normalize the toughness test data, strain data, and non-destructive testing results;

[0049] S32, Feature extraction: Extract the mean value and standard deviation from the absorbed energy, extract the weight index of the failure type from the failure mode, extract the strain peak value and fluctuation range of the key monitoring points from the strain data, and extract the defect distribution density and severity from the non-destructive testing results;

[0050] S33, Model construction: Adopt a scoring model based on a multi-layer neural network, with inputs including features such as absorbed energy, damage mode weight, peak strain, fluctuation range, defect density, and severity;

[0051] S34, Model training: Use historical toughness evaluation data for model training and optimize model parameters through cross-validation;

[0052] S35, Scoring output: Output a comprehensive toughness score, ranging from 0 to 1, and divide the scoring levels.

[0053] Furthermore, the steps for generating the durability evaluation report in step S5 are as follows:

[0054] Step S51, Import strain data, non-destructive testing data, toughness experiment data, and environmental monitoring data into the computer system, and store the data in CSV or HDF5 format;

[0055] Step S52, Perform data cleaning, outlier detection, and normalization on the strain data, non-destructive testing data, toughness experiment data, and environmental monitoring data in step S51, integrate the data, and establish a multi-dimensional data model;

[0056] Step S53, Update the data analysis weekly, and output the strain changes, defect distribution, and toughness analysis results of the arch rib to generate a durability evaluation report.

[0057] Furthermore, the steps for generating the remaining life prediction report in step S6 are as follows:

[0058] Step S61, Input the strain data, defect data, environmental monitoring data, and toughness experiment data of the arch rib to construct a multi-layer neural network model, use the data of the past 5 years as the training set to construct a durability model, regularly update the parameters of the durability model, and output the durability score;

[0059] Step S62, Divide the arch rib into four grades of excellent, good, medium, and poor according to the durability score, and output the durability score heat map and high-risk area distribution map;

[0060] Step S63, Input the durability score, strain data, and environmental monitoring data into the long short-term memory network time series prediction model to output the predicted time point when the durability score drops to poor, predict the remaining life of the arch rib, combine the dynamic change trend of the arch rib durability score, generate a remaining life prediction report, and provide suggestions for long-term reinforcement, monitoring, and maintenance of the bridge.

[0061] Advantages of the present invention

[0062] 1. The durability assessment method of the arch rib of a concrete-filled steel tube arch bridge considering toughness in the present invention introduces toughness indicators into the durability assessment system through impact experiments and fracture analyses, filling the gap in the assessment of the toughness of arch ribs in the existing technology. By comprehensively using strain sensing data, non-destructive testing results, environmental monitoring data, and toughness experiment data, and through integrated analysis using a multi-layer neural network model, the comprehensiveness and accuracy of the assessment results are significantly improved. The long short-term memory network (LSTM) time series prediction model is also used to analyze the dynamic change trend of the durability score of the arch rib, accurately predict the remaining life, and provide a scientific basis for bridge maintenance. By collecting environmental data in real time, a correlation model between temperature, humidity, wind speed, and the durability of the arch rib is established, incorporating environmental impacts into the overall assessment system. According to the assessment results, specific area maintenance suggestions are generated, including reinforcement measures for high-risk areas, adjustment of monitoring density, and environmental protection strategies, enhancing the pertinence and operability of bridge maintenance work.

[0063] 2. The present invention makes up for the deficiencies in result integration and data processing in the existing technology by comprehensively using a variety of detection technologies and combining big data analysis. By providing a remaining life prediction model based on artificial intelligence, the foresight of bridge health management is improved. By dynamically incorporating environmental factors, the assessment results are more in line with the actual service conditions of the bridge, enhancing the reliability of the assessment. By generating specific maintenance suggestions, it provides direct guidance for bridge operation and management, reduces maintenance costs, realizes a comprehensive assessment of the strength, stiffness, and toughness of the arch rib, and comprehensively improves the scientific nature of the assessment system. The present invention can more comprehensively, accurately, and dynamically assess the durability of the arch rib of a concrete-filled steel tube arch bridge, providing a scientific basis for improving the safety and service life of the bridge. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a flowchart of the durability assessment method of the arch rib of a concrete-filled steel tube arch bridge considering toughness in the present invention through impact experiments and fracture analyses. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0065] The present invention will be further explained and illustrated below in conjunction with the drawings and specific embodiments. It should be noted that the specific embodiments do not limit the scope of the rights of the present invention.

[0066] As Figure 1 shown, the present specific embodiment provides a durability assessment method of the arch rib of a concrete-filled steel tube arch bridge considering toughness, including the following steps:

[0067] S1, Sensor Network Arrangement: Install vibrating wire strain gauge sensors of the Geokon Vibrating Wire Strain Gauge model at the key parts of the arch rib at the mid-span position, arch feet, and transverse connection nodes of the concrete-filled steel tube arch bridge, aiming to collect strain and deformation data of the arch rib and provide high-quality data support for subsequent analysis. Connect the vibrating wire strain gauge sensors to the data acquisition system through shielded cables, and the function of the shielded cables is to prevent signal interference. Real-time collect strain data through the data acquisition system and conduct regular calibration; the strain data is recorded at a frequency of once per hour; the time for regular calibration is to conduct initial reference calibration every 24 hours.

[0068] The mid-span position: Monitor the main stress changes at the location of the maximum bending moment of the concrete-filled steel tube arch bridge.

[0069] The arch foot position: As the key support point of the concrete-filled steel tube arch bridge, focus on monitoring the concentration of shear force and bending moment.

[0070] The transverse connection node: As the local stress concentration point of the concrete-filled steel tube arch bridge, evaluate the stress distribution in the transverse connection area.

[0071] The installation steps of the vibrating wire strain gauge sensors are as follows:

[0072] S11, Clean the surface of the installation point to ensure no rust or impurities, and fix the vibrating wire strain gauge sensors using strong glue or bolts; lay the shielded cables for preventing signal interference and connect them to the data acquisition system;

[0073] S12, Initial calibration: Collect the data within 24 hours after installation as the reference value. Parameter settings: The collection frequency is once per hour, and calibration is conducted regularly once every 6 months;

[0074] S13, Data collection and storage: The data is uploaded to the database in real-time through data collection equipment of the Geokon model, and the data storage format: adopts the CSV format or JSON format for easy subsequent analysis.

[0075] S2, Non-destructive testing: Use an ultrasonic detector of the GE USM 36 model and an infrared thermal imager of the FLIR T660 model to conduct non-destructive testing and marking on the internal defects and surface temperature anomalies of the arch rib; the ultrasonic detector is used to identify internal defects such as cracks and cavities. The infrared thermal imager is used to detect surface temperature anomalies. The ultrasonic detector scans the surface of the arch rib at a frequency of 2 MHz to identify internal cracks and cavities; the infrared thermal imager conducts temperature imaging on the surface of the arch rib, records the temperature distribution, and marks the temperature anomaly areas. The specific operation steps are as follows:

[0076] S21, Ultrasonic testing: Apply coupling agent to ensure good contact between the probe and the testing surface. Control the moving speed of the probe at 0.1 m / s. Record the reflection signals and mark the suspected defect areas.

[0077] S22, Infrared thermal imaging: Keep the instrument at a distance of 0.5 - 1 m from the testing surface. Capture the thermal images and mark the areas with abnormal surface temperature.

[0078] S23, Data analysis: Import the ultrasonic data and infrared thermal image data into MATLAB for processing;

[0079] Threshold analysis: Mark as a defect when the intensity of the ultrasonic reflection signal exceeds the set threshold; the threshold is used as the basis for dividing different evaluation levels. For example, when the threshold is set to 0.8 within the scoring range, it is used to distinguish between the "excellent" and "good" levels.

[0080] Thermal image anomaly analysis: Mark the points with abnormal temperature when the temperature exceeds the set range;

[0081] If the temperature exceeds the set allowable range, for example, when the deviation of the temperature from the normal value exceeds ±2°C, or the standard deviation exceeds the set threshold, it is regarded as an abnormal situation.

[0082] Temperature deviation: The temperature exceeds the normal range (for example, the set normal range is ±2°C).

[0083] Standard deviation: The volatility or variation amplitude of the temperature is too large and exceeds the set standard deviation threshold.

[0084] Defect point comparison: Verify the reliability of the defects through the intersection of ultrasonic waves and infrared thermal images;

[0085] Output visual results: Generate scatter plots and thermal images, and mark the defect locations.

[0086] S24, Parameter setting: Conduct a detection once after the completion of bridge construction and conduct regular detections annually during the service period.

[0087] S3, Toughness evaluation: Extract samples from the non-critical stress-bearing areas of the arch rib, and the non-critical stress-bearing areas of the arch rib are the arch rib webs. Use an Instron 8801 impact testing machine to conduct impact tests on the samples. The impact speed of the impact testing machine is set to 5 m / s, and the impact energy is 300 J; the samples are processed into Charpy impact specimens of 10 mm × 10 mm × 55 mm; record the impact absorption energy of the samples and the toughness test data of the failure modes, and the failure modes are brittle fracture and plastic deformation. Combine the strain data and non-destructive testing results in step 1 to establish an integrated toughness scoring model for the arch rib;

[0088] The steps for establishing the integrated scoring model for the arch rib are as follows:

[0089] S31, Data normalization: Normalize the toughness test data, strain data, and non-destructive testing results to ensure the comparability of data in different dimensions.

[0090] S32, Feature extraction: Extract the average value and standard deviation from the absorbed energy, extract the weight index of the failure type from the failure mode, extract the strain peak value and fluctuation range of the key monitoring points from the strain data, and extract the defect distribution density and severity from the non-destructive testing results;

[0091] S33, Model construction: Adopt a scoring model based on a multi-layer neural network, with inputs including features such as absorbed energy, failure mode weight, strain peak value, fluctuation range, defect density, and severity;

[0092] S34, Model training: Use historical toughness evaluation data for model training and optimize the model parameters through cross-validation;

[0093] S35, Score output: Output the comprehensive toughness score, ranging from 0 to 1, and divide the score levels into excellent, good, medium, and poor.

[0094] S4, Environmental parameter monitoring: Arrange a Davis Vantage Pro2 type weather station around the bridge and obtain environmental monitoring data of temperature, humidity, wind speed, and rainfall. Evaluate the impact of external conditions on the durability of the arch rib through environmental monitoring. Use the time series analysis method for the environmental monitoring data and use the Pearson correlation coefficient to analyze the correlation between environmental factors and strain. When conducting extreme condition analysis, examine the impact of two extreme environmental conditions, high temperature and high humidity, on the strain data, and determine the change distribution of the arch rib evaluated by the strain data under high temperature and high humidity conditions. High temperature refers to the temperature value exceeding the 90th percentile of the overall temperature data; high humidity refers to the humidity value exceeding the 85th percentile of the overall humidity data. Conduct correlation analysis with the strain data and non-destructive testing results collected once per hour to generate a time series diagram and a correlation analysis diagram.

[0095] S5, Data integration and durability scoring: Integrate the vibrating wire strain gauge sensor, non-destructive testing, toughness evaluation, and environmental monitoring data, perform multi-dimensional data cleaning, normalization, and modeling processing, and generate a durability evaluation report;

[0096] The steps for generating the durability evaluation report are as follows:

[0097] Step S51: Import the strain data, non-destructive testing data, toughness experiment data, and environmental monitoring data into the computer system. The data is stored in CSV or HDF5 format. The non-destructive testing data is the defect data from ultrasonic testing and infrared thermogram analysis. Step S52: Perform data cleaning, outlier detection, and normalization through MATLAB to remove outliers, map the data to [0,1], integrate the data, and use TensorFlow to build a multi-layer neural network. Its input layer includes strain, defects, toughness indicators, and environmental parameters. The hidden layer has 3 layers and uses the ReLU activation function. The output layer has a single node. The durability score ranges from 0 to 1, and the scoring levels are excellent, good, medium, and poor. The excellent level is set to 0.8 - 1.0, indicating no maintenance required; good is set to 0.6 - 0.8, indicating enhanced monitoring; medium is set to 0.4 - 0.6, indicating defect repair; poor is set to less than 0.4, indicating reinforcement or replacement. The learning rate is set to 0.001, the Batch Size is set to 32, and the Epochs is set to 50. The optimizer is set to Adam.

[0098] Step S53: Update the data analysis weekly, and output the strain changes, defect distribution, and toughness analysis results of the arch rib, and uniformly store them in HDF5 format.

[0099] S6. Remaining life prediction: Based on the prediction model of artificial intelligence, comprehensively score the durability score and toughness state of the arch rib, and generate a remaining life prediction report in combination with the comprehensive score trend;

[0100] The steps for generating the remaining life prediction report described in Step S6 are as follows:

[0101] Step S61: Input the strain data, defect data, environmental monitoring data, and toughness experiment data of the arch rib to build a multi-layer neural network model, use the data of the past 5 years as the training set to build a durability model, regularly update the parameters of the durability model, and output the durability score;

[0102] The steps for building the multi-layer neural network model are as follows:

[0103] (1) Algorithm description

[0104] Input: Durability score time series, environmental parameters.

[0105] Multi-layer neural network model architecture: Multi-layer LSTM. Input layer: Time step (e.g., 30 days), number of features (e.g., score, temperature, etc.). Hidden layer: 2 layers, LSTM, output the future durability score.

[0106] Output: The time when the score time point drops to <0.4.

[0107] (2) Formula

[0108] Loss function: Mean Squared Error (MSE).

[0109] (3) Verification and optimization

[0110] Verification metrics: MSE: Measures the deviation of the predicted score. MAE: Mean Absolute Error.

[0111] Optimization: Hyperparameter tuning (such as the number of hidden layers, Dropout ratio).

[0112] Step S62: Divide the arch ribs into four grades of excellent, good, medium, and poor according to the durability score, and output the durability score heat map and the high-risk area distribution map;

[0113] Step S63: Input the durability score, strain data, and environmental monitoring data into a Long Short-Term Memory (LSTM) time series prediction model to output the predicted remaining life of the arch ribs at the time point when the durability score drops to poor. Combine the dynamic change trend of the arch rib durability score to generate a durability assessment report and a remaining life prediction report, and provide long-term reinforcement, monitoring, and maintenance suggestions for the bridge.

[0114] S7. Feedback and maintenance suggestions: Identify high-risk areas based on the durability assessment report and generate area-specific maintenance suggestions; High-risk areas are where strain exceeds the standard and defects are concentrated. Provide long-term reinforcement, monitoring, and maintenance suggestions for the bridge. Specifically, the reinforcement measures for high-risk areas, increased monitoring density, and periodic environmental protection measures are to be responded to within 1 week; The overall assessment results of the arch ribs are updated monthly.

[0115] Through the synergistic effect of sensor network layout, non-destructive testing, toughness assessment, and artificial intelligence analysis in durability assessment, and combined with environmental monitoring and time series prediction models, the present invention provides a comprehensive and operable solution for the long-term safety and maintenance of arch bridge ribs.

Claims

1. A durability assessment method for the arch rib of a steel tube concrete arch bridge considering toughness, characterized in that: The following steps are involved: S1, sensor network layout: strain gauge sensors are installed on key parts of the arch ribs of the steel tube concrete arch bridge, and the strain gauge sensors are connected to the data acquisition system through shielded cables. The data acquisition system collects strain data in real time and performs regular calibration; S2, non-destructive testing: use ultrasonic detectors and infrared thermal imagers to perform non-destructive testing and marking of internal defects and abnormal surface temperature of arch ribs; S3, toughness assessment: extract samples from the non-critical stress-bearing area of ​​the arch rib, perform impact tests on the samples using an impact testing machine, record the impact absorption energy and toughness test data of the samples, and combine the strain data and non-destructive testing results of step 1 to establish a comprehensive scoring model for arch rib toughness; S4, environmental parameter monitoring: Meteorological stations are arranged around the bridge to obtain environmental monitoring data of temperature, humidity, wind speed and rainfall. The environmental monitoring data are correlated with strain data and non-destructive testing results through time series analysis method to generate time series diagram and correlation analysis diagram; S5, Data Integration and Durability Scoring: Integrate vibrating wire strain gauge sensors, nondestructive testing, toughness assessment and environmental monitoring data, perform multi-dimensional data cleaning, normalization and modeling, and generate a durability assessment report; S6, Remaining life prediction: Based on the prediction model of artificial intelligence, the durability score and toughness status of the arch rib are comprehensively scored, and the remaining life prediction report is generated based on the comprehensive score trend; S7, Feedback and Maintenance Recommendations: Identify high-risk areas based on the durability assessment report and provide recommendations for long-term bridge reinforcement, monitoring and maintenance.

2. The evaluation method according to claim 1, characterized in that: Step S1: The key parts of the arch rib of the steel tube concrete arch bridge are the mid-span position, arch foot and transverse connection node of the steel tube concrete arch rib; the strain gauge sensor is a vibrating wire strain gauge sensor; the strain data is recorded at a frequency of once per hour; and the time for the regular calibration is an initial benchmark calibration every 24 hours.

3. The evaluation method according to claim 1, characterized in that: In step S2, the ultrasonic detector scans the surface of the arch rib at a frequency of 2 MHz to identify internal cracks and cavities; the infrared thermal imager performs temperature imaging on the surface of the arch rib, records the temperature distribution, and marks the temperature abnormality area.

4. The evaluation method according to claim 1, characterized in that: In step S3, the sample is processed into an impact specimen of 10 mm×10 mm×55 mm; the impact speed of the impact testing machine is set to 5 m / s, and the impact energy is 300 J; the failure mode is brittle fracture and plastic deformation.

5. The evaluation method according to claim 1, characterized in that: The steps of establishing the arch rib comprehensive scoring model in step S3 are as follows: S31, data normalization processing: normalizing the toughness test data, strain data and nondestructive testing results; S32, feature extraction: extract the mean value and standard deviation from the absorbed energy, extract the weight index of the damage type from the damage mode, extract the strain peak value and fluctuation range of the key monitoring point from the strain data, and extract the defect distribution density and severity from the non-destructive testing results; S33, model construction: a scoring model based on a multi-layer neural network is used, and the input includes characteristics of absorbed energy, failure mode weight, strain peak, fluctuation range, defect distribution density and severity; S34, training model: using historical resilience assessment data to train the model and optimizing model parameters through cross-validation; S35, score output: output a comprehensive toughness score ranging from 0 to 1, and divide the score into grades.

6. The evaluation method according to claim 1, characterized in that: The steps of generating the durability assessment report in step S5 are as follows: Step S51, storing strain data, nondestructive testing data, toughness test data and environmental monitoring data in CSV or HDF5 format; Step S52, performing data cleaning, outlier detection and normalization processing on the strain data, nondestructive testing data, toughness test data and environmental monitoring data of step S51 through MATLAB, integrating the data to establish a multidimensional data model; Step S53, update the data analysis every week, output the strain change, defect distribution and toughness analysis results of the arch rib, and generate a durability evaluation report.

7. The evaluation method according to claim 1, characterized in that: The steps of generating the remaining life prediction report in step S6 are as follows: Step S61, input the strain data, defect data, environmental monitoring data and toughness test data of the arch rib to build a multi-layer neural network model, and use the data of the past five years as a training set to build a durability model, regularly update the parameters of the durability model, and output a durability score; Step S62, classifying the arch ribs into four grades: excellent, good, medium and poor according to the durability score, and outputting a durability score heat map and a high-risk area distribution map; Step S63, by inputting the durability score, strain data and environmental monitoring data into the long short-term memory network time series prediction model, the output predicts the remaining life of the arch rib at the time point when the durability score drops to a worse point, and combines the dynamic change trend of the arch rib durability score to generate a remaining life prediction report and provide long-term reinforcement, monitoring and maintenance suggestions for the bridge.

Images