Calculation method of bending capacity of steel bar lap joints in underground continuous wall based on big data

Abstract

The present invention discloses a method for calculating the bending capacity of underground continuous wall steel bar lap joints based on big data, which belongs to the technical field of underground engineering design. In order to solve the problem of difficulty in responding to changes in construction site conditions in real time and difficulty in dynamically adjusting construction parameters to adapt to complex construction environments, multi-source data collection and key feature screening are used to effectively improve the quality of the data set and the accuracy of bending capacity prediction. By establishing a physical prediction model for steel bar lap joints, factors such as force transmission and bonding force are comprehensively considered to make the model have good adaptability under different working conditions, dynamically adjust the model parameters to respond to changes in the environment and construction errors in real time, and automatically adjust the weights when monitoring data fluctuations, thereby achieving accurate prediction of bending capacity, effectively preventing potential failure risks, and predicting failure modes and multi-level risk management mechanisms to help construction teams quickly identify high-risk areas and flexibly adjust construction plans.

Description

Technical Field

[0001] The present invention relates to the technical field of underground engineering design, and in particular to a method for calculating the bending resistance of underground continuous wall reinforcement lap joints based on big data. Background Art

[0002] Underground continuous wall is a key component widely used in deep foundation pit support, retaining structure and underground buildings. Its structural performance directly affects the overall stability and safety of the project. In the construction of underground continuous wall, steel lap joint is a common connection method, but due to the complex construction site environment, the bending capacity of steel lap joint is easily affected by many factors, such as lap length, steel bar diameter, concrete strength and construction error. The complexity of these factors makes it difficult for traditional bending capacity assessment methods to adapt to actual engineering needs. Existing technologies usually calculate bending capacity based on empirical formulas or experimental data.

[0003] However, there are still the following problems in actual operation:

[0004] In the existing technology, the traditional model fails to deeply consider the mechanical properties and specific failure mechanisms of the steel bar lap joints, and it is difficult to fully describe the bending resistance of the joints, resulting in poor generalization ability of the model under different construction environments. At the same time, it is unable to flexibly respond to the dynamic changes of the construction site. The influence of the environment and construction errors is often ignored, resulting in the prediction results being inconsistent with the actual situation, affecting construction safety. It is impossible to flexibly adjust the construction parameters according to the on-site feedback, resulting in the mismatch between the construction plan and the actual working conditions on site. Summary of the invention

[0005] The purpose of the present invention is to provide a method for calculating the bending capacity of underground continuous wall steel bar lap joints based on big data to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data, comprising the following steps:

[0007] Data collection, collecting construction parameter data of steel bar lap joints during underground continuous wall construction, the construction parameter data including steel bar lap mode, steel bar diameter, lap length, concrete strength, construction environment and load conditions, wherein the data sources of the construction parameter data include on-site monitoring systems, experimental data and historical engineering data;

[0008] Feature selection, selecting key features related to bending resistance from the collected data, wherein the key features are determined by correlation analysis;

[0009] Establishing a physical model, and establishing a physical prediction model of the steel bar lap joint according to the stress characteristics of the steel bar lap joint, wherein the physical prediction model includes the force transmission of the steel bar lap joint, the bonding force between the concrete and the steel bar, and the stress distribution around the joint;

[0010] Big data regression analysis to establish the regression relationship between the bending resistance of steel bar lap joints and characteristic variables;

[0011] Model calibration and verification: Use the collected data to train the model, calibrate and verify the model with existing experimental data and engineering examples, compare the bending test results in historical engineering cases with the model calculation results, and verify the effectiveness of the model;

[0012] Multivariable dynamic optimization: evaluate the influence of multiple variables on the bending resistance of joints through big data analysis, and modify the calculation model through weight adjustment. The multiple variables include environmental factors and construction errors, and dynamically adjust the parameters of the calculation model in combination with real-time construction monitoring data;

[0013] The result output evaluation is based on the trained physical prediction model to input the characteristic parameters of the underground continuous wall steel bar lap joint, output the predicted bending capacity and conduct a safety assessment to predict the possible failure mode of the joint.

[0014] Furthermore, feature selection specifically includes the following steps:

[0015] Select relevant features closely related to bending resistance from the collected data set, including steel bar lap length, steel bar diameter, joint type, concrete grade, and joint installation environment parameters, and extract and aggregate lap parameter types by statistically analyzing engineering example data;

[0016] Correlation analysis evaluates the correlation between each relevant feature and the bending resistance, screens out the key features that have a significant impact on the bending resistance, and removes invalid features with low correlation;

[0017] Sort the selected key features by importance and determine the weight of each feature in the model.

[0018] Furthermore, the big data regression analysis also includes the following steps:

[0019] Selecting characteristic variables for establishing a regression model from the key characteristics, the characteristic variables including lap length, steel bar diameter, concrete strength and load type parameters, and determining a suitable combination of input characteristic variables through data analysis;

[0020] A linear regression model is established based on the selected characteristic variables, and the relationship between the bending capacity of the steel bar lap joint and each characteristic variable is quantified based on the regression model;

[0021] The performance of the constructed regression model is evaluated to quantify the prediction accuracy of the regression model. If the regression model error is large, the feature variables and model structure are iteratively adjusted until the regression model achieves optimal performance.

[0022] Furthermore, the model calibration verification specifically includes the following steps:

[0023] The collected data is divided into a training set and a validation set. The training set data is used to perform preliminary training on the model, while the validation set data is not involved in model training.

[0024] The model is trained iteratively multiple times based on the training set data, and the model parameters are gradually adjusted. The model fit is optimized by comparing the predicted values ​​with the actual values ​​during the training process.

[0025] Use the validation set to test the model performance and analyze the deviation between the model prediction value and the actual bending resistance. If the deviation is large, continue to adjust the model parameters and structure.

[0026] Furthermore, the multivariable dynamic optimization further includes the following steps:

[0027] Based on big data analysis, the influence of multiple variables on the bending resistance of joints is evaluated. The influence weights of different environmental factors and construction errors on the bending resistance are analyzed through historical data to obtain the influence ranking of multiple variables.

[0028] According to the influence weight of each variable, adjust the parameter weights of different variables in the model;

[0029] Use the real-time monitoring data of on-site construction as model input to modify the model's parameter values;

[0030] By integrating real-time monitoring data, the model automatically adjusts according to the actual environment at the construction site and dynamically outputs the calculation results of the bending capacity of the joint.

[0031] Furthermore, the multivariable dynamic optimization further includes the following steps:

[0032] Real-time monitoring data fusion further includes automatic variable adjustment and feedback learning;

[0033] Automatic variable adjustment is based on real-time environmental monitoring data, including temperature, humidity, and construction error parameters. It detects the changing trend of environmental monitoring data in real time and automatically calculates the dynamic impact weight on bending resistance.

[0034] When the monitoring data fluctuates drastically, the variable adjustment function is automatically triggered to dynamically modify the weights of each variable in the model;

[0035] Feedback learning is based on the bending capacity errors that have occurred in the early stage. It performs error analysis by comparing the predicted values ​​with the actual test data, and records the data of factors with larger errors as feature sets.

[0036] Furthermore, the result output evaluation specifically includes the following steps:

[0037] Based on the trained model, the characteristic parameters of the underground continuous wall reinforcement lap joint are input, and the predicted bending capacity results are output;

[0038] Conducting safety assessment on the bending capacity prediction results, the safety assessment includes bending strength, durability and construction adaptability. If the predicted bending capacity is lower than the design requirements, providing construction adjustment parameters;

[0039] Based on the predicted bending resistance and various characteristic parameters, possible failure modes of the joint are predicted, including slip, shear failure and bending failure.

[0040] Furthermore, the result output evaluation specifically includes the following steps:

[0041] Failure mode prediction further includes multi-level failure risk assessment and failure mode control mechanism;

[0042] Multi-level failure risk assessment constructs a hierarchical failure risk rating system by analyzing the prediction results of the bending capacity of steel bar lap joints under different working conditions;

[0043] Each failure mode is graded according to the probability of occurrence, potential impact range and severity, and a multi-level failure alarm mechanism from low risk to high risk is generated based on the mode classification;

[0044] The failure mode control mechanism monitors the failure mode in real time based on the multi-level failure assessment results. If the risk of the failure mode exceeds the preset threshold, the monitoring frequency is automatically increased and fed back to the construction monitoring system in real time.

[0045] Furthermore, the calculation method of the bending capacity of the underground continuous wall reinforcement lap joint based on big data also includes the use of multi-physics field coupling analysis to verify the bending capacity prediction results. The specific steps include:

[0046] Building a multi-physics coupling model: Based on the construction parameter data, a multi-physics coupling model of continuous wall reinforcement overlap is established. The multi-physics coupling model includes at least mechanical and thermal physical fields.

[0047] Finite element analysis: using finite element analysis (FEA) technology, the response of steel bar lap joints under different physical fields is virtually simulated, and the influence of various physical field parameters on lap failure is determined based on the simulation results;

[0048] Identification of key parameters: Combined with big data technology, statistical analysis of simulation results is performed to identify key parameters affecting overlap and corresponding failure modes;

[0049] Failure mode verification is achieved by comparing the failure mode obtained by multi-physical field coupling analysis with the failure mode predicted by the physical prediction model.

[0050] Furthermore, the failure mode verification specifically includes:

[0051] When the failure mode obtained by multi-physics coupling analysis is inconsistent with the failure mode predicted by the physical prediction model, perform the following steps:

[0052] The process confidence evaluation is performed on the multi-physics field coupling analysis and the physical prediction model prediction respectively. The minimum sample size used in the process confidence evaluation is as follows: in,

[0053] When evaluating process confidence for multi-physics coupled analysis, The minimum sample size of the physical field parameters used in the multi-physics coupling analysis, is the confidence interval of the physical field parameter, , Degree of freedom The quantile of the t distribution, is the deviation of the confidence level of the physical field parameter from 1, is the overall standard deviation of the physical field parameter, is the probability that the physical field parameter lies within the confidence interval, is the total error of the physical field parameters, that is, the standard deviation and confidence interval of the sample mean of the physical field parameters The product of the values; when evaluating the process confidence of the physical prediction model prediction, The minimum sample size of the characteristic parameters of the steel bar lap joint used in the physical prediction model prediction, is the confidence interval of the characteristic parameter, , , is the degree of freedom The quantile of the t distribution, is the deviation value of the confidence level of the characteristic parameter from 1, is the population standard deviation of the characteristic parameter, is the probability that the characteristic parameter is within the confidence interval, is the total error of the characteristic parameter, that is, the standard deviation and confidence interval of the characteristic parameter sample mean The product of values;

[0054] On the basis that the sample size is not less than the minimum sample size, according to the process confidence evaluation result, the failure mode prediction result with higher confidence in the multi-physical field coupling analysis and the physical prediction model prediction is selected as the bending capacity prediction result.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention collects construction parameter data from multiple sources during the construction of underground continuous walls, including on-site monitoring systems, experimental data and historical engineering data, and combines key features determined by correlation analysis to effectively filter out irrelevant factors and optimize the quality of the data set, thereby improving the prediction accuracy of the model for the bending resistance. By establishing a physical prediction model based on the stress characteristics of steel bar lap joints, the force transmission of steel bar lap joints, the bonding force between concrete and steel bars, and the stress distribution around the joints are considered, so that the model can more comprehensively describe the influencing mechanism of the joint's bending resistance, thereby having good adaptability and generalization ability under different working conditions.

[0057] 2. The present invention evaluates the influence weights of different environmental factors and construction errors and dynamically adjusts the model parameters. The model can respond to changes in the on-site environment in real time and integrate real-time monitoring data of on-site construction. The model can adaptively adjust the output results according to current conditions to improve the accuracy of bending capacity prediction and make the prediction results more in line with actual working conditions. The automatic variable adjustment function ensures that the model can identify and respond in time when the monitoring data fluctuates violently and quickly adjust the variable weights. Multivariable dynamic optimization ensures the real-time accuracy of the calculation results of the bending capacity of the joint by identifying and adjusting the influence of the environment and construction errors on the bending capacity, which helps to identify and prevent potential failure risks in the project. The model can monitor and feedback the environment and error changes in real time during the construction process, making the structural design and construction safer and more reliable.

[0058] 3. The present invention uses the failure mode prediction function, and the model can identify potential failure modes and predict the performance of joints under different working conditions in advance, which helps the construction team take measures in advance to avoid the occurrence of failure modes. By grading failure modes according to probability of occurrence, scope of impact and severity, construction personnel can quickly identify and pay attention to high-risk areas, thereby improving the precision and effectiveness of failure management. The failure mode control mechanism monitors the risk situation of failure modes in real time, and dynamically adjusts the monitoring frequency according to multi-level failure assessment results to ensure close attention to high-risk failure modes. The system provides real-time feedback and adjusts parameters to help the construction team flexibly adjust the construction plan according to actual conditions. This dynamic adaptability can timely optimize construction parameters according to the predicted bending resistance and failure mode assessment results, thereby improving the adaptability of the construction process, ensuring the close connection between engineering design and actual construction conditions, and reducing the impact of construction errors on structural safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 The present invention is a flow chart of a method for calculating the bending capacity of underground continuous wall reinforcement lap joints. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] In order to solve the technical problems in the existing technology, the traditional model fails to deeply consider the mechanical properties and specific failure mechanism of the steel bar lap joint, and it is difficult to fully describe the bending resistance of the joint, resulting in poor generalization ability of the model under different construction environments. At the same time, it cannot flexibly respond to the dynamic changes of the construction site. The influence of the environment and construction errors is often ignored, resulting in the prediction results not being consistent with the actual situation, affecting construction safety. The construction parameters cannot be flexibly adjusted according to the feedback from the site, resulting in the mismatch between the construction plan and the actual working conditions on site. Please refer to Figure 1 , the present invention provides the following technical solutions:

[0062] The calculation method of the bending capacity of the steel bar lap joint of underground continuous wall based on big data includes the following steps:

[0063] Data collection, collecting construction parameter data of steel bar lap joints during underground continuous wall construction, the construction parameter data including steel bar lap mode, steel bar diameter, lap length, concrete strength, construction environment and load conditions, wherein the data sources of the construction parameter data include on-site monitoring systems, experimental data and historical engineering data;

[0064] Feature selection, selecting key features related to bending resistance from the collected data, wherein the key features are determined by correlation analysis;

[0065] Establishing a physical model, and establishing a physical prediction model of the steel bar lap joint according to the stress characteristics of the steel bar lap joint, wherein the physical prediction model includes the force transmission of the steel bar lap joint, the bonding force between the concrete and the steel bar, and the stress distribution around the joint;

[0066] Data regression analysis, to establish the regression relationship between the bending capacity of steel bar lap joints and characteristic variables;

[0067] Model calibration and verification: Use the collected data to train the model, calibrate and verify the model with existing experimental data and engineering examples, compare the bending test results in historical engineering cases with the model calculation results, and verify the effectiveness of the model;

[0068] Multivariable dynamic optimization: evaluate the influence of multiple variables on the bending resistance of joints through big data analysis, and modify the calculation model through weight adjustment. The multiple variables include environmental factors and construction errors, and dynamically adjust the parameters of the calculation model in combination with real-time construction monitoring data;

[0069] The result output evaluation is based on the trained physical prediction model to input the characteristic parameters of the underground continuous wall steel bar lap joint, output the predicted bending capacity and conduct a safety assessment to predict the possible failure mode of the joint.

[0070] In the above embodiment, by collecting construction parameter data from multiple sources during the construction of underground continuous walls, covering on-site monitoring systems, experimental data and historical engineering data, and combining key features determined by correlation analysis, irrelevant factors are effectively filtered out and the quality of the data set is optimized, thereby improving the prediction accuracy of the model's bending capacity. By establishing a physical prediction model based on the stress characteristics of steel bar lap joints, considering factors such as the force transfer of steel bar lap joints, the bond between concrete and steel bars, and the stress distribution around the joints, the model can more comprehensively describe the influencing mechanism of the joint's bending capacity, thereby having good adaptability and generalization capabilities under different working conditions.

[0071] In the above embodiment, through the multivariable dynamic optimization module combined with real-time construction monitoring data, the model can evaluate in real time the impact of environmental factors, construction errors and other variables on the bending capacity, automatically adjust the parameter weights, so that the calculation results are consistent with the actual working conditions, and effectively improve the real-time response capability of the model. Through the safety assessment and failure mode prediction in the result output evaluation, the model can predict the bending strength and failure mode of the joint during the construction process, generate graded risk alerts and prevention suggestions, and help construction personnel identify and deal with potential structural risks in advance, ensuring the long-term safety and durability of the underground continuous wall structure.

[0072] In the above embodiment, the entire calculation method is based on automatic analysis of big data and model prediction, which reduces the need for manual calculation and parameter adjustment. Construction personnel can directly rely on the bending resistance results and early warning information output by the model to optimize construction plans and decisions, thereby improving construction efficiency and reducing labor costs.

[0073] Feature selection also includes the following steps:

[0074] Select relevant features closely related to bending resistance from the collected data set, including steel bar lap length, steel bar diameter, joint type, concrete grade, and joint installation environment parameters, and extract and aggregate lap parameter types by statistically analyzing engineering example data;

[0075] Correlation analysis evaluates the correlation between each relevant feature and the bending resistance, screens out the key features that have a significant impact on the bending resistance, and removes invalid features with low correlation;

[0076] Sort the selected key features by importance and determine the weight of each feature in the model.

[0077] In the above embodiment, by screening features closely related to bending resistance and performing correlation analysis to screen out key features, it is ensured that the input data of the model is more accurate and directly related to the bending resistance, thereby improving the prediction accuracy and effectiveness of the model for bending resistance and reducing the interference of irrelevant factors on the calculation results. By sorting the key features in order of importance and determining the weight of each feature in the model, the model can automatically adapt the influence weights of key variables under different working conditions, thereby enhancing the adaptability and robustness of the model under different construction environments and material conditions.

[0078] In the above embodiment, by clarifying the order and weight distribution of key features, the specific impact of each feature on the bending capacity can be more intuitively understood. If the model prediction results deviate from expectations in actual applications, the feature weights can be directly adjusted or replaced to improve the adjustability of the model and its flexibility to actual working conditions, and to screen and aggregate features that have a significant impact on the bending capacity, so that the model can accurately evaluate the bending strength of lap joints. Construction personnel can optimize construction design and quality control in a targeted manner based on the importance of these features in the model, thereby improving overall construction safety and structural durability.

[0079] Big data regression analysis also includes the following steps:

[0080] Selecting characteristic variables for establishing a regression model from the key characteristics, the characteristic variables including lap length, steel bar diameter, concrete strength and load type parameters, and determining a suitable combination of input characteristic variables through data analysis;

[0081] A linear regression model is established based on the selected characteristic variables, and the relationship between the bending capacity of the steel bar lap joint and each characteristic variable is quantified based on the regression model;

[0082] The performance of the constructed regression model is evaluated to quantify the prediction accuracy of the regression model. If the regression model error is large, the feature variables and model structure are iteratively adjusted until the regression model achieves optimal performance.

[0083] In the above embodiment, by selecting characteristic variables for establishing a regression model from key features and performing reasonable combination optimization, the model input is made more accurate, ensuring that the regression model can accurately capture the relationship between bending resistance and various influencing factors, thereby improving the accuracy of bending resistance prediction. By establishing a linear regression model that can quantify the relationship between bending resistance and characteristic variables, it can be flexibly adapted according to the input of different characteristic variables under different working conditions and material conditions.

[0084] In the above embodiment, the performance of the constructed regression model is evaluated, and the prediction accuracy of the model is quantified to ensure that the calculation results of the model meet the accuracy requirements; if the model error is large, the feature variable combination or the model structure is iteratively adjusted. Through this automatic optimization process, the model can be continuously improved in practical applications, thereby improving the reliability of the prediction. Through the optimal selection and combination of feature variables, irrelevant or low-impact data is removed, data redundancy is avoided, and the efficiency of data processing and regression model calculation is improved, providing support for real-time analysis and rapid decision-making, and reducing computing costs.

[0085] In the above embodiment, the relationship between the bending capacity and the key variables is quantified by the regression model, so that construction personnel can scientifically evaluate the bending capacity of the steel bar lap joints and intuitively understand the specific influence of the lap length, steel bar diameter, concrete strength and load type on the bending capacity, thereby supporting construction design and on-site parameter adjustment to ensure structural safety and economy.

[0086] Model calibration verification also includes the following steps:

[0087] The collected data is divided into a training set and a validation set. The training set data is used to perform preliminary training on the model, while the validation set data is not involved in model training.

[0088] The model is trained iteratively multiple times based on the training set data, and the model parameters are gradually adjusted. The model fit is optimized by comparing the predicted values ​​with the actual values ​​during the training process.

[0089] Use the validation set to test the model performance and analyze the deviation between the model prediction value and the actual bending resistance. If the deviation is large, continue to adjust the model parameters and structure.

[0090] In the above embodiment, the model is gradually adjusted according to the difference between the predicted value and the actual value during the training process, and the model is continuously optimized through multiple iterative training. The model parameters are continuously optimized during the training process, thereby improving the model fit, making it more in line with the actual situation of the bending capacity, improving the model performance, detecting the model performance deviation through the validation set, and quickly identifying and adjusting the parameters and model structure with large errors, thereby improving the debugging efficiency. When the deviation is large, the model structure or parameter combination can be iteratively adjusted until the model reaches the optimal performance, avoiding unnecessary redundant training steps, saving development time and computing resources, and the calibrated and verified model has a higher prediction accuracy, which can provide more reliable bending capacity assessment results for construction. Construction personnel can scientifically formulate construction plans based on the prediction results of the model, and reasonably select construction materials and process parameters, thereby improving the bending capacity of the underground continuous wall steel lap joints and improving the safety and durability of the overall project.

[0091] Multivariable dynamic optimization also includes the following steps:

[0092] Based on big data analysis, the influence of multiple variables on the bending resistance of joints is evaluated. The influence weights of different environmental factors and construction errors on the bending resistance are analyzed through historical data to obtain the influence ranking of multiple variables.

[0093] According to the influence weight of each variable, adjust the parameter weights of different variables in the model;

[0094] Use the real-time monitoring data of on-site construction as model input to modify the model's parameter values;

[0095] Integrating real-time monitoring data, the model automatically adjusts according to the actual environment at the construction site and dynamically outputs the calculation results of the bending resistance of the joint;

[0096] Real-time monitoring data fusion further includes automatic variable adjustment and feedback learning;

[0097] Automatic variable adjustment is based on real-time environmental monitoring data, including temperature, humidity, and construction error parameters. It detects the changing trend of environmental monitoring data in real time and automatically calculates the dynamic impact weight on bending resistance.

[0098] When the monitoring data fluctuates drastically, the variable adjustment function is automatically triggered to dynamically modify the weights of each variable in the model;

[0099] Feedback learning is based on the bending capacity errors that have occurred in the early stage. It performs error analysis by comparing the predicted values ​​with the actual test data, and records the data of factors with larger errors as feature sets.

[0100] In the above embodiment, by evaluating the influence weights of different environmental factors and construction errors and dynamically adjusting the model parameters, the model can respond to changes in the on-site environment in real time, integrate the real-time monitoring data of on-site construction, and the model can adaptively adjust the output results according to the current conditions to improve the accuracy of the bending capacity prediction, so that the prediction results are more in line with the actual working conditions. The automatic variable adjustment function ensures that the model can identify and respond in time when the monitoring data fluctuates violently, and quickly adjust the variable weights. This process ensures that the model can still output stable and reliable bending capacity calculation results when responding to rapid changes in the environment and construction conditions, which improves the flexibility and stability of the model in complex on-site environments.

[0101] In the above embodiment, feedback learning automatically identifies factors with large errors by comparing predicted values ​​with actual detection data and records them as feature sets, thereby continuously optimizing the feature selection and weight distribution of the model, enhancing the generalization ability of the model under diversified data conditions, enabling the model to accurately predict bending capacity in more environments, and expanding the applicability of the model. The dynamically optimized model can not only output real-time bending capacity results through real-time monitoring data fusion and feedback learning, but also provide automated monitoring and decision-making support for construction. Construction personnel can obtain dynamic evaluation results of joint bending capacity in real time, and adjust construction plans in time according to system prompts, thereby optimizing the construction process and reducing potential structural risks.

[0102] In the above embodiment, multivariable dynamic optimization ensures the real-time accuracy of the calculation results of the bending capacity of the joint by identifying and adjusting the influence of the environment and construction errors on the bending capacity, which helps to identify and prevent potential failure risks in the project. The model can monitor and feedback the environment and error changes in real time during the construction process, making the structural design and construction safer and more reliable.

[0103] The result output evaluation also includes the following steps:

[0104] Based on the trained model, the characteristic parameters of the underground continuous wall reinforcement lap joint are input, and the predicted bending capacity results are output;

[0105] Conducting safety assessment on the bending capacity prediction results, the safety assessment includes bending strength, durability and construction adaptability. If the predicted bending capacity is lower than the design requirements, providing construction adjustment parameters;

[0106] Based on the predicted bending resistance and various characteristic parameters, possible failure modes of the joint are predicted, including slip, shear failure and bending failure;

[0107] Failure mode prediction further includes multi-level failure risk assessment and failure mode control mechanism;

[0108] Multi-level failure risk assessment constructs a hierarchical failure risk rating system by analyzing the prediction results of the bending capacity of steel bar lap joints under different working conditions;

[0109] Each failure mode is graded according to the probability of occurrence, potential impact range and severity, and a multi-level failure alarm mechanism from low risk to high risk is generated based on the mode classification;

[0110] The failure mode control mechanism monitors the failure mode in real time based on the multi-level failure assessment results. If the risk of the failure mode exceeds the preset threshold, the monitoring frequency is automatically increased and fed back to the construction monitoring system in real time.

[0111] In the above embodiment, safety assessment is performed on the bending capacity prediction results, including bending strength, durability and construction adaptability, to ensure that the predicted bending capacity meets the design requirements. If the prediction result is lower than the safety standard, the system will actively provide construction adjustment suggestions to ensure that the bending capacity of the steel bar lap joints during construction meets the design standards, effectively improving the safety and reliability of the structure. Through the failure mode prediction function, the model can identify potential failure modes and predict the performance of the joints under different working conditions in advance, which helps the construction team take measures in advance to avoid failure modes. This forward-looking assessment enhances the initiative of engineering safety management and reduces the possibility of structural failure.

[0112] In the above embodiment, the multi-level failure risk assessment is based on the bending capacity prediction results under different working conditions to establish a hierarchical failure risk level system. By classifying the failure modes according to the probability of occurrence, scope of impact and severity, a hierarchical alarm mechanism from low risk to high risk is generated. Construction personnel can quickly identify and pay attention to high-risk areas, thereby improving the precision and effectiveness of failure management. The failure mode control mechanism monitors the risk situation of the failure mode in real time, and dynamically adjusts the monitoring frequency according to the multi-level failure assessment results to ensure close attention to high-risk failure modes. When the risk exceeds the preset threshold, the system automatically increases the monitoring frequency and feeds back to the construction monitoring system, so that construction personnel can respond quickly and take preventive measures, significantly improving the real-time monitoring level of the construction site.

[0113] In the above embodiment, the system helps the construction team to flexibly adjust the construction plan according to the actual situation by providing real-time feedback and adjusting parameters. This dynamic adaptability can timely optimize the construction parameters according to the predicted bending resistance and failure mode assessment results, improve the adaptability of the construction process, ensure the close connection between the engineering design and the actual construction situation, and reduce the impact of construction errors on structural safety. By establishing a failure alarm and real-time feedback mechanism, it can timely warn and take preventive measures before the failure risk occurs, reducing the maintenance and reinforcement costs caused by joint failure. In addition, by dynamically adjusting the construction parameters, material waste and construction delays are effectively reduced, thereby reducing the overall project risk and operating costs and improving the project's economic benefits.

[0114] On the basis of the above-mentioned embodiment, the method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data also includes using multi-physical field coupling analysis to verify the bending capacity prediction results. The specific steps include:

[0115] Building a multi-physics coupling model: Building a multi-physics coupling model for continuous wall reinforcement overlap based on construction parameter data. The multi-physics coupling model includes at least mechanical and thermal physical fields.

[0116] Finite element analysis: using finite element analysis (FEA) technology, the response of steel bar lap joints under different physical fields is virtually simulated, and the influence of various physical field parameters on lap failure is determined based on the simulation results;

[0117] Identification of key parameters: Combined with big data technology, statistical analysis of simulation results is performed to identify key parameters affecting overlap and corresponding failure modes;

[0118] Failure mode verification is achieved by comparing the failure mode obtained by multi-physical field coupling analysis with the failure mode predicted by the physical prediction model.

[0119] In the above embodiment, multi-physical field coupling analysis is used to verify the bending capacity prediction result. During the multi-physical field coupling analysis, a multi-physical field coupling model is built and finite element analysis is introduced. The influence of each physical field parameter on the lap failure is determined in the simulation, and the key parameters affecting the lap and the corresponding failure mode are identified. The parameters with the greatest influence are taken as the key parameters, and the failure mode is determined according to the specific requirements of the key parameters. This is used to verify the bending capacity prediction result obtained by the physical prediction model. If the two are consistent, it means that the verification is passed. If the two are inconsistent, it is necessary to further analyze the cause to determine whether adjustment is needed or other necessary measures are taken to ensure that the steel bar lap joint meets the safety requirements.

[0120] Based on the above embodiment, the failure mode verification specifically further includes:

[0121] When the failure mode obtained by multi-physics coupling analysis is inconsistent with the failure mode predicted by the physical prediction model, perform the following steps:

[0122] The process confidence evaluation is performed on the multi-physics field coupling analysis and the physical prediction model prediction respectively. The minimum sample size used in the process confidence evaluation is as follows: in,

[0123] When evaluating process confidence for multi-physics coupled analysis, The minimum sample size of the physical field parameters used in the multi-physics coupling analysis, is the confidence interval of the physical field parameter, , Degree of freedom The quantile of the t distribution, is the deviation of the confidence level of the physical field parameter from 1, is the overall standard deviation of the physical field parameter, is the probability that the physical field parameter is within the confidence interval, is the total error of the physical field parameters, that is, the standard deviation and confidence interval of the sample mean of the physical field parameters The product of the values; when evaluating the process confidence of the physical prediction model prediction, The minimum sample size of the characteristic parameters of the steel bar lap joint used in the physical prediction model prediction, is the confidence interval of the characteristic parameter, , , is the degree of freedom The quantile of the t distribution, is the deviation value of the confidence level of the characteristic parameter from 1, is the population standard deviation of the characteristic parameter, is the probability that the characteristic parameter is within the confidence interval, is the total error of the characteristic parameter, that is, the standard deviation and confidence interval of the characteristic parameter sample mean The product of the values; on the basis that the sample size is not less than the minimum sample size, according to the process confidence evaluation result, the failure mode prediction result with higher confidence in the multi-physical field coupling analysis and the physical prediction model prediction is selected as the bending capacity prediction result.

[0124] In the above embodiment, a method for handling when the failure mode obtained by the multi-physical field coupling analysis is inconsistent with the failure mode predicted by the physical prediction model is provided, that is, by performing confidence assessment on both, the failure mode prediction result with higher confidence is determined as the bending capacity prediction result, rather than simply drawing a conclusion. In the process of confidence assessment, the sample size used in the assessment is also limited, that is, not less than the minimum sample size calculated by the above algorithm, so as to ensure the reliability of the confidence assessment result and avoid misjudgment, thereby improving the accuracy and credibility of the bending capacity prediction result.

[0125] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data, characterized in that: The following steps are involved: Data collection, collecting construction parameter data of steel bar lap joints during underground continuous wall construction, the construction parameter data including steel bar lap mode, steel bar diameter, lap length, concrete strength, construction environment and load conditions, wherein the data sources of the construction parameter data include on-site monitoring systems, experimental data and historical engineering data; Feature selection, selecting key features related to bending resistance from the collected data, wherein the key features are determined by correlation analysis; Establishing a physical model, and establishing a physical prediction model of the steel bar lap joint according to the stress characteristics of the steel bar lap joint, wherein the physical prediction model includes the force transmission of the steel bar lap joint, the bonding force between the concrete and the steel bar, and the stress distribution around the joint; Big data regression analysis to establish the regression relationship between the bending resistance of steel bar lap joints and characteristic variables; Model calibration and verification: Use the collected data to train the model, calibrate and verify the model with existing experimental data and engineering examples, compare the bending test results in historical engineering cases with the model calculation results, and verify the effectiveness of the model; Multivariable dynamic optimization: evaluate the influence of multiple variables on the bending resistance of joints through big data analysis, and modify the calculation model through weight adjustment. The multiple variables include environmental factors and construction errors, and dynamically adjust the parameters of the calculation model in combination with real-time construction monitoring data; The result output evaluation is based on the trained physical prediction model to input the characteristic parameters of the underground continuous wall steel bar lap joint, output the predicted bending capacity and conduct a safety assessment to predict the possible failure mode of the joint.

2. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1 is characterized in that: Feature selection also includes the following steps: Select relevant features closely related to bending resistance from the collected data set, including steel bar lap length, steel bar diameter, joint type, concrete grade, and joint installation environment parameters, and extract and aggregate lap parameter types by statistically analyzing engineering example data; Correlation analysis evaluates the correlation between each relevant feature and the bending resistance, screens out the key features that have a significant impact on the bending resistance, and removes invalid features with low correlation; Sort the selected key features by importance and determine the weight of each feature in the model.

3. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1 is characterized in that: Big data regression analysis also includes the following steps: Selecting characteristic variables for establishing a regression model from the key characteristics, the characteristic variables including lap length, steel bar diameter, concrete strength and load type parameters, and determining a suitable combination of input characteristic variables through data analysis; A linear regression model is established based on the selected characteristic variables, and the relationship between the bending capacity of the steel bar lap joint and each characteristic variable is quantified based on the regression model; The performance of the constructed regression model is evaluated to quantify the prediction accuracy of the regression model. If the regression model error is large, the feature variables and model structure are iteratively adjusted until the regression model achieves optimal performance.

4. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1 is characterized in that: Model calibration verification also includes the following steps: The collected data is divided into a training set and a validation set. The training set data is used to perform preliminary training on the model, while the validation set data is not involved in model training. The model is trained iteratively multiple times based on the training set data, and the model parameters are gradually adjusted. The model fit is optimized by comparing the predicted values ​​with the actual values ​​during the training process. Use the validation set to test the model performance and analyze the deviation between the model prediction value and the actual bending resistance. If the deviation is large, continue to adjust the model parameters and structure.

5. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1 is characterized in that: Multivariable dynamic optimization also includes the following steps: Based on big data analysis, the influence of multiple variables on the bending resistance of joints is evaluated. The influence weights of different environmental factors and construction errors on the bending resistance are analyzed through historical data to obtain the influence ranking of multiple variables. According to the influence weight of each variable, adjust the parameter weights of different variables in the model; Use the real-time monitoring data of on-site construction as model input to modify the model's parameter values; By integrating real-time monitoring data, the model automatically adjusts according to the actual environment at the construction site and dynamically outputs the calculation results of the bending capacity of the joint.

6. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 5 is characterized in that: Multivariable dynamic optimization also includes the following steps: Real-time monitoring data fusion further includes automatic variable adjustment and feedback learning; Automatic variable adjustment is based on real-time environmental monitoring data, including temperature, humidity, and construction error parameters. It detects the changing trend of environmental monitoring data in real time and automatically calculates the dynamic impact weight on bending resistance. When the monitoring data fluctuates drastically, the variable adjustment function is automatically triggered to dynamically modify the weights of each variable in the model; Feedback learning is based on the bending capacity errors that have occurred in the early stage. It performs error analysis by comparing the predicted values ​​with the actual test data, and records the data of factors with larger errors as feature sets.

7. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1, characterized in that: The result output evaluation also includes the following steps: Based on the trained model, the characteristic parameters of the underground continuous wall reinforcement lap joint are input, and the predicted bending capacity results are output; Conducting safety assessment on the bending capacity prediction results, the safety assessment includes bending strength, durability and construction adaptability. If the predicted bending capacity is lower than the design requirements, providing construction adjustment parameters; Based on the predicted bending resistance and various characteristic parameters, possible failure modes of the joint are predicted, including slip, shear failure and bending failure.

8. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 7 is characterized in that: The result output evaluation also includes the following steps: Failure mode prediction further includes multi-level failure risk assessment and failure mode control mechanism; Multi-level failure risk assessment constructs a hierarchical failure risk rating system by analyzing the prediction results of the bending capacity of steel bar lap joints under different working conditions; Each failure mode is graded according to the probability of occurrence, potential impact range and severity, and a multi-level failure alarm mechanism from low risk to high risk is generated based on the mode classification; The failure mode control mechanism monitors the failure mode in real time based on the multi-level failure assessment results. If the risk of the failure mode exceeds the preset threshold, the monitoring frequency is automatically increased and fed back to the construction monitoring system in real time.

9. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data as claimed in claim 1, characterized in that: It also includes the use of multi-physics field coupling analysis to verify the bending capacity prediction results. The specific steps include: Building a multi-physics coupling model: Building a multi-physics coupling model for continuous wall reinforcement overlap based on construction parameter data. The multi-physics coupling model includes at least mechanical and thermal physical fields. Finite element analysis: using finite element analysis technology, the response of steel bar lap joints under different physical fields is virtually simulated, and the influence of various physical field parameters on lap failure is determined based on the simulation results; Identification of key parameters: Combined with big data technology, statistical analysis of simulation results is performed to identify key parameters affecting overlap and corresponding failure modes; Failure mode verification is achieved by comparing the failure mode obtained by multi-physical field coupling analysis with the failure mode predicted by the physical prediction model.

10. The method for calculating the bending capacity of underground continuous wall reinforcement lap joints based on big data according to claim 9, characterized in that: Failure mode verification also includes: When the failure mode obtained by multi-physics coupling analysis is inconsistent with the failure mode predicted by the physical prediction model, perform the following steps: The process confidence evaluation is performed on the multi-physics field coupling analysis and the physical prediction model prediction respectively. The minimum sample size used in the process confidence evaluation is as follows: in, When evaluating process confidence for multi-physics coupled analysis, The minimum sample size of the physical field parameters used in the multiphysics coupling analysis, is the confidence interval of the physical field parameter, , Degree of freedom The quantile of the t distribution, is the deviation of the confidence level of the physical field parameter from 1, is the overall standard deviation of the physical field parameter, is the probability that the physical field parameter lies within the confidence interval, is the total error of the physical field parameters, that is, the standard deviation and confidence interval of the sample mean of the physical field parameters The product of the values; when evaluating the process confidence of the physical prediction model prediction, The minimum sample size of the characteristic parameters of the steel bar lap joint used in the physical prediction model prediction, is the confidence interval of the characteristic parameter, , , is the degree of freedom The quantile of the t distribution, is the deviation value of the confidence level of the characteristic parameter from 1, is the population standard deviation of the characteristic parameter, is the probability that the characteristic parameter is within the confidence interval, is the total error of the characteristic parameter, that is, the standard deviation and confidence interval of the characteristic parameter sample mean The product of the values; on the basis that the sample size is not less than the minimum sample size, according to the process confidence evaluation result, the failure mode prediction result with higher confidence in the multi-physical field coupling analysis and the physical prediction model prediction is selected as the bending capacity prediction result.

Images