Search

Lightweight materials with better resistance to sliding wear are prominent candidates for automobile brake drums, clutch pads and cylinder block applications to facilitate fuel economy. This attempt is reserved to cater to materials with higher tribological quality needs. Less dense foamy pumice stone particles were involved in three different percentages (5, 10, and 15 wt%) to reinforce lightweight AZ31 Mg alloy. A stir-assisted squeeze casting technique was pursued to process the composite and refine the grain structure. A phase detection, elemental mapping and microstructure study were done through X-ray diffraction (XRD), energy dispersive spectroscopy (EDS), and scanning electron microscopy (SEM), respectively. An experimental dry sliding wear scrutiny was administered using a pin-on-disc apparatus by considering: (i) ambient and elevated temperature environments and (ii) three different levels of loads. The results reveal a significant drop in wear loss and a frictional coefficient for 15% pumice-loaded composite than the base alloy. Post-wear examination acknowledges the fact that the ambient temperature wear is governed by adhesive-abrasive wear and high temperature is by abrasive wear mechanisms. Worn-out scrutiny authenticates the presence of oxide layers and their role in lubrication. A comparative study with previous works upholds the novel magnesium composite is the right candidate for the mentioned automobile applications.

The present investigation focuses on the microstructural behavior and mechanical properties of Aluminium alloy 2024 reinforced with SiC nanoparticles by applying the Friction Stir Processing (FSP) technique. Taguchi L9 orthogonal array was used to find the optimal process parameters. The experiment used the expected best process parameters, which confirmed the predicted highest value for the mechanical characteristic. Traverse speed, axial load, and rotating speed are the main factors affecting aluminum metal matrix composite. Hence, this experiment examined the effects of five different tool spinning speeds, specifically 800, 900, 1000, 1100, and 1200 rpm. Weld samples are fabricated wherein the stir zone contains SiC nanoparticles. The results of the study indicate that the optimal welding rotational speed of 1000 rpm has a significant impact on the microstructure, wear rate, and microhardness. Specifically, it was observed that the stirred zone exhibited enhanced wear resistance compared to other zones. The study's findings revealed that after implementing friction stir processing (FSP), a well-defined shear zone was observed on the advancing side of the stir zone. Additionally, it is found that the uniform dispersion and strong bonding of SiC particles with an aluminium matrix further enhance wear resistance.

The incorporation of nanomaterials (NMs) in concrete will produce the utmost properties. Nevertheless, investigational work takes a great deal of time with efforts to measure the strength of concrete. Thus, the process of estimating the properties of cement paste can be accelerated by employing soft machine learning (SML) techniques. Individual machine learning approaches (MLAs) like random forest (RF), decision tree (DT), and support vector machine (SVM) were utilized. These individual MLAs models then ensemble with two approaches namely bagging and boosting. The outcome of the model depicts a robust performance by showing significant correlations (R2) as compared to individual MLAs. Data points with 205 having contribution factors (e.g., curing time (days), water to binder (w/b), micro-silica (%), and nano-silica (%) were used as input parameters, and compressive strength as output for MLAs modeling. To ensure that each model is as accurate as possible, cross-validation with K-folding and statistical error analysis (i.e., MAE, MAE, and RMSE) were utilized. Thus, interpretable approaches with AdaBoost give a strong correlation of R2 ranging from 0.85 to 0.90 respectively with fewer errors. Out of which DT with ensemble Adaboost and RF give a robust performance. Moreover, the results of the cross-validation show that the response was accurate, with fewer errors. Therefore, the ensemble model represents a robust correlation. Therefore, its usage with individual ML would enhance the model's performance. In addition, the graphical user interface (GUI) is created by training a model that can predict the values of the desired outputs when the necessary input parameters are provided. It simplifies the procedure and provides a handy tool for employing the model's capabilities in the civil engineering domain.

This research used gene expression programming (GEP) and multi expression programming (MEP) to determine the compressive strength (CS) of alkali-activated material (AAM) to compare and develop more reliable genetic algorithm-based prediction models. To learn more about how raw ingredients affect and interact with the CS of AAM, a SHapley Additive exPlanations (SHAP) analysis was conducted. A comprehensive dataset containing 676 points with fifteen influential parameters was formulated from the previously published literature. According to this study, considering the impact of 15 input variables, both genetic algorithms produced results close to the experimental CS (retrieved from the literature). When the performance of the GEP and MEP models were compared, it was found that the MEP model, with an R2 of 0.86, performed better than the GEP model, with an R2 of 0.82. The assessment of the statistical parameters of generated models revealed that the MEP model was more effective. Additionally, SHAP analysis revealed that slag content, followed by the specimen's age, sodium silicate, and curing temperature, showed a positive correlation with CS of AAM, which were the most important parameters. The results also revealed the importance of chemical contents, i.e., CaO, SiO2, Al2O3, of FA and slag on the CS of AAM. The built models might be used to compute the CS of AAMs with varying input parameter values, minimizing the effort, time, and cost of unnecessary lab tests. Furthermore, the outcomes of the SHAP study might help researchers and the industry determine the quantity or composition of raw ingredients when producing AAMs.

Alkali-activated materials (AAMs) have gained significant attention in recent years due to their potential for sustainable construction and outstanding strength and durability characteristics. However, their wide application is constrained by their weak behavior under flexure and tensile loading, leading to brittleness. This study examines the formulation of an alkali-activated fiber-reinforced mortar, using fly ash and hydrated lime as main precursors, and investigates the capacity when reinforced with polyvinyl alcohol (PVA) and polypropylene (PP) fibers. The performance was measured through experimental investigation in terms of flowability, setting time, compressive strength (CS), water absorption, flexure strength (FS), and load-displacement curve under flexure loading. The durability was assessed through acid and sulfate attack tests after 120 days. The findings showed that adding both types of fibers improved the mechanical and durability performance of the one-part AAM. The highest CS of 44.95 MPa and 42.04 MPa and FS of 7.8 MPa and 6.9 MPa were achieved for PVA and PP fibers at a 1.5% fiber content. The FS was increased by 18.94% and 28.2% for PP1.5 and PVA1.5 fibers respectively. The control sample, PP 1.5 and PVA 1.5 recorded 31.1%, 12.5%, and 11% loss in CS after the acid attack test, respectively, after 120 days. A reduction of 15.3% and 13.3% was recorded in flow diameter with the addition of PP and PVA fibers at 2%. Overall, mortars reinforced with PVA fibers have shown better flowability, mechanical strength, and chemical resistance than PP fibers.

Rice Husk ash (RHA) utilization in concrete as a waste material can contribute to the formation of a robust cementitious matrix with utmost properties. The strength of HPC when subjected to compression test is determined by the combination and quantity of the materials used in its production. Thus, making its mixed design process challenging and ambiguous. The objective of this research is to forecast the strength of HPC containing RHA, by using a diverse range of machine learning techniques, including both individual and ensemble learners such as bagging and boosting. This study will cause significant implications for sustainable construction practices by facilitating the development of an efficient and effective method for forecasting the strength of HPC. Individual machine learning (ML) algorithms are incorporated with ensemble methods such as bagging, adaptive boosting, and random forest algorithms. These ensemble techniques is use to create twenty different sub-models. Afterward, these sub-models is train and optimized for achieving the best possible value for R2. The sub-models were further fine-tuned to obtain the best value for R2. In order to assess or evaluate the quality, reliability, and generalizability of the test data, the K-Fold cross-validation method is utilized. Furthermore, the index for measuring the statistical performance of models is use to validate and compare the assessment of ensemble models with individual models. The findings indicate that using bagging and boosting techniques enhances the prediction accuracy of individual or weak models, with minimum errors and an R2 value > 0.92 is achieved using bagging with decision tree and random forest. In general, the performance of the model is optimized by using ensemble learner methods in machine learning (ML).

This study aims to develop various machine learning (ML) models to investigate the self-healing capacity of engineered cementitious composites (ECC) and to evaluate the effect of input parameters (raw materials and crack width before the healing process) on output parameters (crack width after the healing process) via Shapley additive explanations (SHAP) analysis. The combination of the individual and ensemble ML models has been introduced to check and compare the accuracy level towards the prediction of crack width after the healing process to access the most suitable model for this application. The support vector machine (SVM) from the individual, while XGBoost (XGB) and random forest (RF) from ensemble ML models have been investigated for prediction purposes. As per the obtained results, the RF model outperforms both the SVM and XGB algorithms in forecasting the fracture width after the healing process for the selected ECC in concrete material. The performance indicator, such as the coefficient of determination (R2), was reported as 0.97 for RF, 0.96 for XGB, and 0.93 for the SVM model. Statistical results and k-fold cross-validation for the employed ML models also confirm their legitimacy. It was also noted from the result of the SHAP analysis that the significant contribution was for the crack width before healing towards the prediction of crack width after the healing process. Furthermore, the ML models can also be utilized to anticipate the healing ability of the other ECC, such as soda glass powder, marble powder, and bagasse ash.

This study used three artificial intelligence-based algorithms – adaptive neuro-fuzzy inference system (ANFIS), artificial neural networks (ANNs), and gene expression programming (GEP) – to develop empirical models for predicting the compressive strength (CS) and slump values of fly ash-based geopolymer concrete. A database of 245 CS and 108 slump values were established from the published literature, where 17 significant parameters were chosen as input variables for the development of models. The algorithms were trained and tested using statistical measures including Nash-Sutcliffe efficiency, root-squared error, root-mean-square error, relative-root-mean-square error, mean absolute error, correlation coefficient, and regression coefficient. The comparison results showed that the GEP model was superior to the ANFIS and ANN models in terms of R-value, R2, and RMSE for both CS and slump prediction. The R-value for the CS models was 0.94 (GEP), 0.92 (ANFIS), and 0.91 (ANN), while for the slump it was 0.96 (GEP), 0.91 (ANFIS), and 0.90 (ANN). Moreover, the performance index factor values for slump and CS were found 0.03 and 0.029 for GEP-models and 0.036, 0.030 for ANFIS-models and 0.035 and 0.034 for ANN-models respectively. The sensitivity and parametric analysis were also performed for GEP-developed model. Results demonstrate that the GEP model generates more accurate prediction for the slump and CS of fly ash-based geopolymer after being rigorously trained and its hyperparameters optimized.

To investigate the effect of eggshell powder on the water-absorption capacity of cement mortar, this research employed experimental testing followed by machine learning (ML) modeling techniques. On cement mortar specimens, the eggshell powder was partially substituted for cement and sand in concentrations ranging from 0% to 15% with a 2.5% increment, and water absorption tests were conducted. Experimental tests were used to generate the dataset, which was then used to train ML algorithms and construct ML-based prediction models. The integration of eggshell powder lowered the water absorption capacity of cement mortar, as determined by the test findings. The highest reduction in water absorption was seen at 5% eggshell powder as a cement replacement, resulting in up to 41.7% decrease, and 7.5% eggshell powder as a sand replacement, resulting in up to 45.8% decrease in water absorption compared to the control sample. In addition, the constructed ML models exhibited good agreement with the actual results and may be utilized to determine the water absorption of cement mortar incorporating eggshell powder. Nevertheless, based on the R2 value, the disparity between experimental and predicted results, and error analysis, it was determined that the random forest model was more accurate than the decision tree model.

This study evaluates the compressive strength (C–S) of nano-silica-based fiber-reinforced concrete (NS-FRC) by using advanced machine learning (ML) individual and ensembled techniques. The employed advanced ML approaches used for the analysis are Support Vector Machine (SVM), Multi-Layer Perceptron (MLP) and eXtreme Gradient Boosting (XGB). Furthermore, the level of accuracy for the employed advanced algorithms is also evaluated by the k-fold cross-validation technique. Statistical checks, i.e., root mean square error (RMSE), mean absolute error (MAE) and mean absolute percent error (MAPE), are also applied to validate the performance of algorithms. Sensitivity analysis is also made to explore the influence of input parameters on the C–S of NS-FRC. Among all, the XGB technique is found most effective for an accurate C–S prediction of NS-FRC. In XGB model, the coefficient of determination (R2) is 0.95, which is comparatively more than that of SVM (0.90) and MLP (0.90). The MAE value of XGB algorithm is 3.3 MPa which is lower than that of SVM (4.8 MPa) and MLP (4.5 MPa). In addition, RMSE value is also less for XGB algorithm (3.8 MPa) as compared to that of SVM (5.5 MPa) and MLP (5.9 MPa). Furthermore, the employed XGB models exhibited highest R2 of 0.95 as compared to the models reported in the available literature. The sensitivity analysis revealed that the nano-silica influenced the C–S of NS-FRC by 7%. Moreover, discussion reveals that nano-silica in concrete can have several benefits, such as improved microstructure, enhanced strength, prolonged durability, reduced cement content, and less carbon emission.