Prediction Of Bond Strength Research Articles

Hyperparameter optimization for interfacial bond strength prediction between fiber-reinforced polymer and concrete

This study proposed an intelligent method based on hyperparameter optimization to predict the strength of the interfacial bond between fiber-reinforced polymer (FRP) and concrete. Because ordinary machine learning models extract features manually, this requires additional time and causes errors in parameter selection. High-precision machine learning model selection and automatic hyperparameter optimization can help overcome these limitations. Comparing eight different machine learning models (i.e., LReg, KNN, LR, MLP, BRR, SVR, LightGBM, and CBR), CatBoost was selected as the primary model for the hyperparameter optimization. CatBoost showed the best performance with the R2 of 0.9394 and MAPE of 1.21%. According to the prediction results, the hyperparameter optimization reduced the dispersion degree by 90 %. In general, the machine learning model works better than the existing models in terms of the coefficient of determination, root mean square error, and coefficient of variation. Furthermore, the model enhanced by the hyperparameter optimization was better than the selected CatBoost model, which indicates that hyperparameter optimization is a reliable approach to improve the accuracy of the model.

Investigation on the effect of Zn interlayer and process parameters optimisation in the production of Al-SS bimetallic castings

The effect of Zn interlayer on the stainless steel insert in the fabrication of Aluminium-Stainless Steel (Al-SS) Bimetallic casting (BmC) is reported. The design of Experiments was conducted with pouring temperature (680°C–730°C), insert preheat temperature (100°C–400°C) and insert thickness (1–3 mm) as influencing process parameters to produce Al-SS BmC. The maximum bond strength of 25.33 MPa was achieved when the experimental conditions were 705°C pouring temperature, 250°C insert temperature and 2 mm insert thickness. Using Response Surface Methodology (RSM), the interaction between the process parameters has been studied and a second-order polynomial equation was derived for maximising the bond strength. The predicted maximum bond strength using RSM for the above experimental conditions is 26.75 MPa. The output of RSM was given to the Genetic Algorithm (GA) to identify the maximum bond strength at forbidden parametric values. The maximum bond strength predicted using GA by interpolation is 26.87 MPa. However, the three process parametric conditions are 707°C, 277°C, and 1.86 mm respectively for achieving the maximum bond strength in the given Al-SS BmC’s. Experimental validation of the above conditions was conducted and the results obtained confirmed the predicted bond strength value using the GA technique. Spiral morphology due to the formation of Zn10Fe3 intermetallics was observed at the Zn-SS interface when the pouring temperature increased to 730°C and this might have led to the reduced bond strength of BmC than the bond strength achieved at 705°C and 707°C. The formation of the Al-rich and Fe-rich solid solutions as reaction layer between Al and SS without the micro gap around 705°C may contribute to the significant increase in bond strength of produced BmC with different insert temperatures and insert thickness.

Bond-slip behaviours between preplaced aggregate concrete (PAC) and steel bars

In this article, 54 specimens with diverse types and strength grades of concrete and various embedded lengths and diameter of steel bars were prepared and the bond-slip behaviours between preplaced aggregate concrete (PAC) and steel bars was evaluated by pull-out tests. Then, a bond stress-slip constitutive relationship was proposed on the basis of the thick-walled cylinder model. Finally, the accuracy of the proposed bond-slip constitutive relationship was verified by finite element analysis (FEA) simulation. The results indicated that the bond stress-slip curves could be roughly divided into four sections. The bond strength for PAC under different concrete strength grades were all superior to those for normal concrete (NC). The proposed theoretical model provided a good prediction of bond strength for PAC with reasonable parameters selection. The simulated bond stress-slip curves for PAC were basically in accordance with the testing results which illustrated that the proposed bond-slip constitutive relationship was reasonable. In the future, the accuracy of the evaluation of the bearing capacity of precast concrete structures composed of PAC will be significantly improved by introducing the proposed bond-slip constitutive model.

Enhancing Sustainability of Corroded RC Structures: Estimating Steel-to-Concrete Bond Strength with ANN and SVM Algorithms

The bond strength between concrete and corroded steel reinforcement bar is one of the main responsible factors that affect the ultimate load-carrying capacity of reinforced concrete (RC) structures. Therefore, the prediction of accurate bond strength has become an important parameter for the safety measurements of RC structures. However, the analytical models are not enough to estimate the bond strength, as they are built using various assumptions and limited datasets. The machine learning (ML) techniques named artificial neural network (ANN) and support vector machine (SVM) have been used to estimate the bond strength between concrete and corroded steel reinforcement bar. The considered input parameters in this research are the surface area of the specimen, concrete cover, type of reinforcement bars, yield strength of reinforcement bars, concrete compressive strength, diameter of reinforcement bars, bond length, water/cement ratio, and corrosion level of reinforcement bars. These parameters were used to build the ANN and SVM models. The reliability of the developed ANN and SVM models have been compared with twenty analytical models. Moreover, the analyzed results revealed that the precision and efficiency of the ANN and SVM models are higher compared with the analytical models. The radar plot and Taylor diagrams have also been utilized to show the graphical representation of the best-fitted model. The proposed ANN model has the best precision and reliability compared with the SVM model, with a correlation coefficient of 0.99, mean absolute error of 1.091 MPa, and root mean square error of 1.495 MPa. Researchers and designers can apply the developed ANN model to precisely estimate the steel-to-concrete bond strength.

Prediction of FRP-concrete interfacial bond strength based on machine learning

Externally bonding fiber reinforced polymer (FRP) to concrete structures is an effective way to enhance the mechanical performance of concrete structures. Many equations have been proposed to predict the interfacial bond strength for FRP-concrete structures but have limited accuracy due to the complexity of the bond behavior. This study proposes to formulate the FRP-concrete interfacial bond strength based on machine learning (ML) methods, which have emerged as a promising alternative to achieve high prediction accuracy in high-dimension problems. To this end, a database containing 1,375 FRP-concrete direct shear test specimens that failed due to interfacial debonding was established. The database was improved using an unsupervised isolation forest that identified and eliminated anomalous data, and was then used to train six ML models, namely artificial neural networks (ANN), support vector machine, decision tree, gradient boosting decision tree, random forest, and XGboost algorithms, to predict the FRP-concrete interfacial bond strength. The ML predictive models showed higher accuracy than 16 existing equations in the literature. The XGBoost model showed the highest accuracy, and its coefficient of variation was 54% lower than the existing equation with the highest accuracy among those considered. The ANN model was used to perform a parametric study on the influencing parameters, and a new equation was generated to predict the interfacial bond strength, considering the key influencing parameters. The equation enables interpretation of the ML models. The study combines ML models and traditional physical models to achieve a novel, interpretable ML method for predicting FRP-concrete interfacial bond strength.

An improved model for predicting bond strength of deformed rebars in concrete with stirrups

To ensure the long-term stability and safety of concrete structures, it is critical to accurately analyze the stirrup confinement to bond performances of deformed bars in concrete. However, it remains a challenging task. As a result of this study, a bond prediction model incorporating the internal pressure, failure plane characteristics, and stirrup confinement is developed as a solution. The internal pressure of the rebar/concrete interface at various stages of comprehensive concrete cracking is determined with the aid of the SIFs-based crack propagation criterion, where the key parameters are evaluated by weight function theory and finite element model. In a comprehensive comparison, the proposed model accurately predicts both stirrup confinement and failure modes. A validation study is performed on 125 experimental results, with the integral absolute error (IAE) of 12.1% for the proposed model, which is significantly lower than other models.

Performance assessment of bamboo bond strength in cement–fly ash mortar

The novelty of this study is the combined experimental and statistical approach, which is showcased to predict and optimise bamboo–mortar bond strength, incorporating the effects of various material factors. The design of experiments method was used to quantify the impact of multiple factors affecting the bond strength. Firstly, different experimental pull-out tests were conducted in the laboratory to determine the effect of various factors (water/cement ratio, curing duration, cement/sand ratio, fly ash content and bamboo treatment) on bond strength development. Three failure patterns were observed – bamboo failure, bond failure and mortar failure. Secondly, using a two-level full factorial design method, different statistical models were formulated. These models were developed based on the surface treatment of the bamboo (model M1 (untreated) and model M2 (treated and untreated)). The effects of single factors and their interactions on bond strength development were quantified. The developed models were validated for the prediction of bond strength based on different factors. The developed statistical models were found to predict and simulate the experimental behaviour of bond strength with sufficient accuracy.

Probabilistic bond strength prediction between the corroded reinforcing bars and concrete considering the concrete strength and non-uniform corrosion

The bond strength considering concrete strength and non-uniformly corrosion was tested by twelve beam-type specimens, and the corrosion was characterized by the 3D scanner. According to the K-S test, the cross-sectional area of the corroded bar follows the lognormal distribution at the 5% significance level, where its mean and standard deviation linearly correlate with the mass loss rate. The probabilistic bond strength is predicted by the Monte-Carlo method, incorporating the uncertainties of the cross-sectional area of rebars and the concrete strength. It is indicated that the failure probability can be decreased by decreasing mass loss rate and increasing concrete strength.

Development of probabilistic FRP-to-concrete bond strength models for externally-bonded reinforcement on grooves: Bayesian approach

• FRP-to-concrete bond strength models are developed for EBROG joints. • An experimental database containing 304 test samples is collected. • Bayesian inference is adopted for probabilistic model calibration. • Analytical expressions are derived for probabilistic model prediction. • Traditional models are compared to illustrate the superior of proposed ones. Recent years have witnessed the superiorities of a grooving technique, i.e., the externally-bonded reinforcement on grooves (EBROG), over the traditional technique, i.e., externally-bonded reinforcement (EBR), in terms of strengthening reinforced-concrete members by fiber-reinforced polymer (FRP) composites. Similar to EBR, the prediction of interfacial bond strength is critical for the EBROG type of FRP-to-concrete joints. There have been a series of experiments conducted to study the bond strength of EBROG joints, however, the corresponding prediction models (especially those capable of uncertainty quantification) are rarely available. Therefore, this paper aims to develop probabilistic models able to both predict the bond strength of EBROG joints and quantify the associated uncertainties. Firstly, an experimental database containing 304 bond strength tests for EBROG joints with longitudinal grooves is collected. Secondly, two widely-accepted deterministic bond strength models are selected as candidates to be calibrated by including a model correction factor. Thirdly, the factor is expressed by power and exponential functions of the influencing variables, meanwhile the deterministic model parameters are replaced with probabilistic ones. Finally, the probabilistic models are determined by means of Bayesian inference and Markov chain Monte Carlo. Moreover, an approximate analytical expressions are derived to achieve the probabilistic bond strength prediction. The experimental database is used to investigate the model performances, and the results show that: 1) the developed models can predict the FRP-to-concrete bond strength of EBROG joints with sufficient predictive accuracies and reasonable uncertainty quantifications; 2) the groove geometry has significant influence on the model correction factor; and 3) the developed probabilistic models is only applicable to EBROG joints with longitudinal grooves, but not to ones with transverse/diagonal grooves.

Bond strength prediction model of defective grout materials in half-grouted sleeve connections under uniaxial and cyclic loadings

The grouted sleeve connection is one of the most popular approaches to connect the reinforcing bars in the prefabricated structures. Insufficient grouting defects of sleeve connections significantly could decrease the bond strength of the grout materials; therefore, increase the failure risk of connection joints. To obtain a prediction model for the bond strength between reinforcing bars and grout materials under uniaxial tensile and cyclic loading conditions is essential for the design and construction. The stress transfer path of the defective grout material and its restraint effect of the sleeve were analyzed in terms of the uniaxial tensile and cyclic loading tests on ninety-nine specimens of half-grouted sleeves with defects. The mechanism was illustrated through the tests and the performance indicators of the connection joints and the deterioration of the grout bonding strength were quantified as a model. The defect-induced instability in the sleeve restraint system is one of the most significant reasons which leads to the bond deterioration in grout materials with defects. Under the same conditions, a concentrated and laterally distributed defect will cause a more severe deterioration than a uniformly dispersed defect. Compared to uniaxial tension, with the defect rate over 30%, the cyclic loading significantly amplifies the adverse effect of defects on the bond performance; however, if a defect rate is less than 20%, the loading methods will not influence on the bond performance of the grout materials. A prediction model for the bond strength is proposed with consideration of the restraint of the sleeve and the properties of the grout itself. Comparing with a large amount of test data, the predicted results can match well with the test results when the defect rate is less than 40%. The proposed work can provide the guidance for the performance assessment of defective grouted sleeves in both seismic and non-seismic zones.

Machine Learning Models for Predicting Bond Strength of Deformed Bars in Concrete

First Name is required invalid characters Last Name is required invalid characters Email Address is required Invalid Email Address Invalid Email Address

Bond-slip behavior of self-consolidating rubberized concrete

• Study of the bond-slip behavior in self-consolidating rubberized concrete (SCRC) • Overview of the available studies on bond-slip behavior in SCRC. • Results show, the bond strength depends on the diameter of reinforcement bar and rubber content. • Proposition of the analytical model for predicting bond strength and corresponding slip. • Ultimate bond strength as a function of the reinforcement diameter, compressive strength and rubber content. A sustainable solution was found for the use of waste material such as used tires by partially replacing the natural aggregate in the concrete with recycled rubber particles. This resulted in improved ductility, but at the expense of reduced strength and stiffness. To further investigate the behavior of rubberized concrete, this paper investigates the bond-slip behavior of reinforcement in self-consolidating rubberized concrete. A total of 27 cubic specimens with reinforcement bars and embedded confinement were constructed. Two parameters affecting the bond-slip behavior were investigated: the reinforcement bar diameter (∅12, ∅16, ∅20 mm), the content of rubber particles in the self-consolidating concrete (0%, 10%, 15%), and silica fume (0%, 5%). From the experimental results, it can be concluded that both a higher rubber content and a larger reinforcement diameter have a negative effect on the bond coefficient, reducing it by up to 28.87% and 18.78%, respectively, while the bond strength decreases by 33% for a rubber replacement of 15%. In the case of rubberized concrete, a more ductile behavior is observed due to the rubber content. Based on the obtained results, a parametric analysis was performed and an analytical prediction model was proposed.

Bond-slip behavior of steel reinforced UHPC under flexure: Experiment and prediction

Ultra-high performance concrete (UHPC) is an advanced class of concrete materials that show superior mechanical and durability performance. While several studies have investigated the bond-slip behavior of steel reinforced UHPC using pullout tests, very limited information is available on the bond-slip behavior of reinforced UHPC from beam-type tests, which produce flexural stress states commonly seen in structural members. This study experimentally examines the local bond-slip behavior of reinforced UHPC using beam-end tests. The test variables include UHPC material designs (one low-cement, green UHPC mix and one typical UHPC mix), fiber volumes (0%, 0.5%, and 1.0%), cover thickness (16 mm–51 mm), steel bar sizes (16 mm–32 mm), and test ages (3 days and 50 days). A total of thirty-two tests are conducted. While the non-fiber UHPC specimens fail in a brittle manner due to splitting failure, all fiber-reinforced UHPC specimens exhibit confined splitting failure and ductile bond-slip responses. The peak bond strength is found to exhibit a linear relationship with the cover-thickness-to-bar-diameter ratio and the square root of UHPC compressive strength. Furthermore, a local bond-slip model is developed, which includes a new bond-strength prediction method that predicts the experimental results with a mean absolute error of 7.6%. Finally, this local bond-slip model is combined with a partial interaction model to explore the global bond stress variations under different embedment lengths and flexural states. Equations are proposed to predict the development length that can lead to reinforcement yielding with minimized bonded length or minimized global slip.

Prediction and uncertainty quantification of ultimate bond strength between UHPC and reinforcing steel bar using a hybrid machine learning approach

The composite action of the reinforcing bars in the UHPC involves complex and nonlinear mechanisms. Inadequate knowledge of their interaction may lead to insufficient bond resistance in the concrete structures. The behavior of reinforcing steel bars in UHPC is generally determined by conducting pull-out tests. However, these tests need amounts of time and cost. A soft computing technique is an alternative approach that exterminates the need to conduct pull-out tests. This study presents the application of artificial intelligence technique in predicting ultimate bond strength (UBS) between ultrahigh-performance concrete (UHPC) and steel reinforcement bars and performs uncertainty quantification of such values. A total of 333 pull-out test data points were collected. A hybrid machine learning (ML) model, improved eliminate particle swamp optimization hybridized artificial neural network (IEPANN) was developed to predict the UBS between UHPC and reinforcing bars. The efficacy of the IEPANN was evaluated against other ML models such as particle swamp optimization hybridized artificial neural network (PANN), support vector regression (SVR), and multiple linear regression (MLR). The IEPANN model was benchmarked against three design codes, including CEB-FIP, AS3600, EC2, and four empirical models. The result obtained based on R = 0.973, MAE = 1.880, RMSE = 2.595, MAPE = 7.708, and RI = 0.944 shows that the IEPANN model can predict UBS between UHPC and reinforcing steel bars with acceptable accuracy. The uncertainty of the UBS was quantified with geometrical and material configuration randomness based on Monte Carlo simulation. The simulation result shows that the yield strength of steel rebar is the most sensitive, while the embedment length is the least sensitive to the bond strength. Systematic assessment of uncertainties in UBS in the design of reinforced concrete members may lead to a higher level of confidence and enhance construction reliability.

Bond strength prediction of timber-FRP under standard and acidic/alkaline environmental conditions based on gene expression programming

Bond strength prediction of timber-FRP under standard and acidic/alkaline environmental conditions based on gene expression programming

Predicting Bond Strength between FRP Rebars and Concrete by Deploying Gene Expression Programming Model.

Rebars made of fiber-reinforced plastic (FRP) might be the future reinforcing material, replacing mild steel rebars, which are prone to corrosion. The bond characteristics of FRP rebars differ from those of mild steel rebars due to their different stress-strain behavior than mild steel. As a result, determining the bond strength (BS) qualities of FRP rebars is critical. In this work, BS data for FRP rebars was investigated, utilizing non-linear capabilities of gene expression programming (GEP) on 273 samples. The BS of FRP and concrete was considered a function of bar surface (Bs), bar diameter (db), concrete compressive strength (fc′), concrete-cover-bar-diameter ratio (c/d), and embedment-length-bar-diameter ratio (l/d). The investigation of the variable number of genetic parameters such as number of chromosomes, head size, and number of genes was undertaken such that 11 different models (M1–M11) were created. The results of accuracy evaluation parameters, namely coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) imply that the M11 model outperforms other created models for the training and testing stages, with values of (0.925, 0.751, 1.08) and (0.9285, 0.802, 1.11), respectively. The values of R2 and error indices showed that there is very close agreement between the experimental and predicted results. 30 number chromosomes, 9 head size, and 5 genes yielded the optimum model. The parametric analysis revealed that db, c/d, and l/d significantly affected the BS. The FRP rebar diameter size is greater than 10 mm, whereas a l/d ratio of more than 12 showed a considerable decrease in BS. In contrast, the rise in c/d ratio revealed second-degree increasing trend of BS.

Neural network model for bond strength of FRP bars in concrete

Interest in FRP composite bars as reinforcement to concrete has increased over the years as it showed solutions to the drawbacks of steel such as its corrosion issues and vulnerability when employed in adverse environmental conditions. However, it is still not widely incorporated as a replacement to conventional steel primarily due to the complexity of its bond strength mechanism. This, therefore, imposes the need to establish a comprehensive relationship for the bond property of the FRP reinforced concrete. This paper developed a novel Artificial Neural Network (ANN) bond strength prediction model for FRP reinforced concrete using 184 hinged beam database from various existing experiments. From series of simulations performed, the model N 7-10-1 with ten nodes in the hidden layer appeared to be the best fit with the experimental results yielded the most favorable performance among other existing models. From the parametric analysis conducted, the compressive strength of the FRP reinforced concrete has proved to be the most dominant parameter in evaluating its bond behavior as determined by relative importance of 17.82%. Overall, the proposed ANN model has demonstrated the best prediction for FRP bond strength in comparison to previous studies and code equations.

The Efficiency of Hybrid Intelligent Models in Predicting Fiber-Reinforced Polymer Concrete Interfacial-Bond Strength

Fiber-reinforced polymer (FRP) has several benefits, in addition to excellent tensile strength and low self-weight, including corrosion resistance, high durability, and easy construction, making it among the most optimum options for concrete structure restoration. The bond behavior of the FRP-concrete (FRPC) interface, on the other hand, is extremely intricate, making the bond strength challenging to estimate. As a result, a robust modeling framework is necessary. In this paper, data-driven hybrid models are developed by combining state-of-the-art population-based algorithms (bald eagle search (BES), dynamic fitness distance balance-manta ray foraging optimization (dFDB-MRFO), RUNge Kutta optimizer (RUN)) and artificial neural networks (ANN) named “BES-ANN”, “dFDB-MRFO -ANN”, and “RUN-ANN” to estimate the FRPC interfacial-bond strength accurately. The efficacy of these models in predicting bond strength is examined using an extensive database of 969 experimental samples. Compared to the BES-ANN and dFDB-MRFO models, the RUN-ANN model better estimates the interfacial-bond strength. In addition, the SHapley Additive Explanations (SHAP) approach is used to help interpret the best model and examine how the features influence the model’s outcome. Among the studied hybrid models, the RUN-ANN algorithm is the most accurate model with the highest coefficient of determination (R2 = 92%), least mean absolute error (0.078), and least coefficient of variation (18.6%). The RUN-ANN algorithm also outperformed mechanics-based models. Based on SHAP and sensitivity analysis method, the FRP bond length and width contribute more to the final prediction results.

Probabilistic models for predicting bond strength of externally bonded FRP-to-concrete joints based on Bayesian inference

Nowadays, externally bonded (EB) fiber-reinforced polymer (FRP) sheets/plates have been recognized as a major strengthening approach for aged concrete structures. In order to evaluate the capacities of strengthened structures, a model capable of predicting the bond strength accurately and quantifying the corresponding uncertainty reasonably is necessary for EB FRP-to-concrete joints. In this study, a probabilistic approach is proposed to calibrate existing deterministic bond strength models to improve the accuracies and quantify the uncertainties of model predictions. A large and extensive experimental database containing 1,228 single-/double-lap shear test samples is firstly collected. Three widely-accepted deterministic bond strength models presented by design codes are then selected as candidates of which the predictive performances are evaluated through the database. After that, the potential influencing variables are added to the candidate bond strength models and the deterministic model parameters are replaced with the probabilistic ones. The probabilistic model calibration is proposed to be accomplished based on Bayesian inference which can yield probability distributions associated with the model parameters and predictions by means of the Markov chain Monte Carlo algorithm. Furthermore, the Kolmogorov-Smirnov test is adopted to study the probabilistic distributions of parameters and predictions of the calibrated models, as well as several statistical evaluation metrics are used to verify the performance of the probabilistic models. The proposed probabilistic models are capable of predicting the bond strength of EB FRP-to-concrete joints with improved predictive accuracies and reasonable uncertainty quantifications.

Application of neuro-fuzzy estimation in prediction of shear bond strength between concrete layers through the efficient laser roughness analyzer

To ensure the monolithic behavior of reinforced concrete composite parts, the bond strength at the interface between concrete layers cast at various ages must be high. Reinforced concrete composite members include precast beams with cast-in-place slabs, bridge decks strengthened by adding a new concrete layer, and repair and strengthening of existing concrete structural members by adding a new concrete layer. in this study, the shear bond strength (SBS) of reinforcing steel on Portland cement and a hybrid cement including slag and Portland cement activated with sodium carbonate is investigated. The pull-out test was used to determine SBS; also a subsequent test data is evaluated using ANFIS, as well as surface classification utilizing a laser roughness analyzer designed particularly to assess the roughness of the concrete substrate. As a result, this research tried to create an in situ non-destructive approach for assessment of concrete surfaces and its influence on shear bond strength measurement of the concrete layers. All test data is analyzed using an adaptive neural fuzzy inference system (ANFIS) to classify the different input variables for determining the shear bond strength between concrete layers using the mean and maximum surface roughness (SR) height parameters Ra and Rt in X and Y direction. ANFIS was used to optimize the process based on five processing parameters. Skillful prediction could play a pivotal role in the optimal conditions during laser cutting process. Based on results, laser speed is the most influential on the Ra in X and Y direction (RMSE: 0.3255, RMSE: 0.6869, respectively). The most influential parameter on the Rt in X direction is laser power (RMSE: 1.5611), while the most influential parameter on the Rt in Y direction is laser speed (RMSE: 2.0781), resulting that the roughness of the substrate surface highly affects the shear bond strength of concrete interfaces.