The joint roughness coefficient(JRC)is a key parameter in the assessment of mechanical properties and the stability of rock masses.This paper presents a novel approach to JRC evaluation using a genetic algorithm-optim...The joint roughness coefficient(JRC)is a key parameter in the assessment of mechanical properties and the stability of rock masses.This paper presents a novel approach to JRC evaluation using a genetic algorithm-optimized backpropagation(GA-BP)neural network.Conventional JRC evaluations have typically depended on two-dimensional(2D)and three-dimensional(3D)parameter calculation methods,which fail to fully capture the nonlinear relationship between the complex surface morphology of joints and their roughness.Our analysis from shear tests on eight different joint types revealed that the strength and failure characteristics of the joints not only exhibit directional dependence but also positively correlate with surface dip angles,heights,and back slope morphological features.Subsequently,five simple statistical parameters,i.e.average dip angle,median dip angle,average height,height coefficient of variation,and back slope feature value(K),were utilized to quantify these characteristics.For the prediction of JRC,we compiled and analyzed 105 datasets,each containing these five statistical parameters and their corresponding JRC values.A GA-BP neural network model was then constructed using this dataset,with the five morphological characteristic statistics serving as inputs and the JRC values as outputs.A comparative analysis was performed between the GA-BP neural network model,the statistical parameter method,and the fractal parameter method.This analysis confirmed that our proposed method offers higher accuracy in evaluating the roughness coefficient and shear strength of joints.展开更多
Evaluating the performance of existing concrete structures is essential in civil engineering,with compressive strength serving as an indicator of performance.Non-destructive testing(NDT)techniques are commonly employe...Evaluating the performance of existing concrete structures is essential in civil engineering,with compressive strength serving as an indicator of performance.Non-destructive testing(NDT)techniques are commonly employed due to their cost-effectiveness and the ability to assess structural integrity without causing damage.However,NDT methods often yield less accurate results than destructive testing(DT),which,although highly reliable,is costly and invasive.To address this limitation,recent research has focused on developing predictive models that correlate DT and NDT outcomes using machine learning techniques.This study explores the application of Support Vector Machine(SVM)models,enhanced with optimization techniques,to improve prediction accuracy.Experimental concrete practical samples,ranging from M10 to M40 grade,were prepared and tested at 14 and 28 days of curing,totaling 126 laboratory specimens.Additionally,231 field samples were collected from a 20-year-old structure to reflect in situ conditions.The performance of SVM was improved using optimization algorithms such as Bayesian Optimization and Genetic Algorithms(GA).Among various kernel functions tested,the Gaussian non-linear kernel proved most effective in modeling the complex relationship between NDT and DT results.The SVM model optimized using Bayesian methods and a Gaussian kernel achieved superior performance,with a high coefficient of determination(R²=0.9771)and significantly lower error metrics,including Mean Squared Error(MSE),Root Mean Squared Error(RMSE),and Mean Absolute Error(MAE).Bayesian-optimized SVM with a Gaussian kernel offers a highly accurate and practical tool for predicting compressive strength from NDT data,enhancing decision-making in structural assessment.展开更多
基金funded by the National Natural Science Foundation of China(Grant Nos.42472345,52325905,42407202).
文摘The joint roughness coefficient(JRC)is a key parameter in the assessment of mechanical properties and the stability of rock masses.This paper presents a novel approach to JRC evaluation using a genetic algorithm-optimized backpropagation(GA-BP)neural network.Conventional JRC evaluations have typically depended on two-dimensional(2D)and three-dimensional(3D)parameter calculation methods,which fail to fully capture the nonlinear relationship between the complex surface morphology of joints and their roughness.Our analysis from shear tests on eight different joint types revealed that the strength and failure characteristics of the joints not only exhibit directional dependence but also positively correlate with surface dip angles,heights,and back slope morphological features.Subsequently,five simple statistical parameters,i.e.average dip angle,median dip angle,average height,height coefficient of variation,and back slope feature value(K),were utilized to quantify these characteristics.For the prediction of JRC,we compiled and analyzed 105 datasets,each containing these five statistical parameters and their corresponding JRC values.A GA-BP neural network model was then constructed using this dataset,with the five morphological characteristic statistics serving as inputs and the JRC values as outputs.A comparative analysis was performed between the GA-BP neural network model,the statistical parameter method,and the fractal parameter method.This analysis confirmed that our proposed method offers higher accuracy in evaluating the roughness coefficient and shear strength of joints.
文摘Evaluating the performance of existing concrete structures is essential in civil engineering,with compressive strength serving as an indicator of performance.Non-destructive testing(NDT)techniques are commonly employed due to their cost-effectiveness and the ability to assess structural integrity without causing damage.However,NDT methods often yield less accurate results than destructive testing(DT),which,although highly reliable,is costly and invasive.To address this limitation,recent research has focused on developing predictive models that correlate DT and NDT outcomes using machine learning techniques.This study explores the application of Support Vector Machine(SVM)models,enhanced with optimization techniques,to improve prediction accuracy.Experimental concrete practical samples,ranging from M10 to M40 grade,were prepared and tested at 14 and 28 days of curing,totaling 126 laboratory specimens.Additionally,231 field samples were collected from a 20-year-old structure to reflect in situ conditions.The performance of SVM was improved using optimization algorithms such as Bayesian Optimization and Genetic Algorithms(GA).Among various kernel functions tested,the Gaussian non-linear kernel proved most effective in modeling the complex relationship between NDT and DT results.The SVM model optimized using Bayesian methods and a Gaussian kernel achieved superior performance,with a high coefficient of determination(R²=0.9771)and significantly lower error metrics,including Mean Squared Error(MSE),Root Mean Squared Error(RMSE),and Mean Absolute Error(MAE).Bayesian-optimized SVM with a Gaussian kernel offers a highly accurate and practical tool for predicting compressive strength from NDT data,enhancing decision-making in structural assessment.
