Predicting Compressive Strength of Blast Furnace Slag and Fly Ash Based Sustainable Concrete Using Machine Learning Techniques: An Application of Advanced Decision-Making Approaches

Experimental analysis and gene expression programming optimization of sustainable concrete containing mineral fillers

Rapid urbanization has led to a high demand for concrete, causing a significant depletion of vital natural resources, notably river sand, which is crucial in the manufacturing process of concrete. As a result, there is a growing need for environmentally sustainable alternatives to fine aggregate in concrete. Quarry dust (QD) has evolved as a viable and ecologically friendly substitute in response to this demand. In the past, limited experimental investigations and only conventional modeling techniques were used to promote sustainable mineral fillers in concrete. This study proposed a robust soft computing technique using gene-expression programming (GEP) to enhance the usability of sustainable alternatives in concrete. Initially, an experimental study was carried out to examine the feasibility and mechanical characteristics of concrete made from materials including quarry dust and superplasticizer as a partial replacement for fine aggregate. Ten mixed proportions with various proportions (0%, 20%, 40%, and 60%) of quarry dust were used to make M15 and M20 grades of concrete. A series of experimental tests, such as workability, compressive strength (CS), and tensile strength (TS), were conducted to examine the fresh and hardened properties of modified concrete. The established database from the experimental investigations was then used to develop machine learning (ML) models using GEP. The outcomes of the GEP models were validated by comparing them with multi-linear regression (MLR) models and using various statistical metrics such as root mean squared error (RMSE), performance index (PI), correlation coefficient (R), and external validation methods. Finally, sensitivity analysis was performed to investigate the influence of ingredients such as mineral fillers, superplasticizers, and others on the mechanical properties of concrete. To enhance the practical usage of the study, a graphical user interface (GUI) was also created. The study revealed that 40% of the replacement of fine aggregates with mineral filler and superplasticizer shows the optimum properties. GEP models outperformed MLR, achieving R² values of 0.96 in CS and 0.92 in TS, compared to MLR's lower values of 0.85 in CS and 0.81 in TS. The proposed GEP equations and user-friendly GUI can be used to develop the pre-mix design of concrete using quarry dust and superplasticizers.

Advanced machine learning algorithms to evaluate the effects of the raw ingredients on flowability and compressive strength of ultra-high-performance concrete

The estimation of concrete characteristics through artificial intelligence techniques is come out to be an effective way in the construction sector in terms of time and cost conservation. The manufacturing of Ultra-High-Performance Concrete (UHPC) is based on combining numerous ingredients, resulting in a very complex composite in fresh and hardened form. The more ingredients, along with more possible combinations, properties and relative mix proportioning, results in difficult prediction of UHPC behavior. The main aim of this research is the development of Machine Learning (ML) models to predict UHPC flowability and compressive strength. Accordingly, sophisticated and effective artificial intelligence approaches are employed in the current study. For this purpose, an individual ML model named Decision Tree (DT) and ensembled ML algorithms called Bootstrap Aggregating (BA) and Gradient Boosting (GB) are applied. Statistical analyses like; Determination Coefficient (R2), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) are also employed to evaluate algorithms' performance. It is concluded that the GB approach appropriately forecasts the UHPC flowability and compressive strength. The higher R2 value, i.e., 0.94 and 0.95 for compressive and flowability, respectively, of the DT technique and lesser error values, have higher precision than other considered algorithms with lower R2 values. SHAP analysis reveals that limestone powder content and curing time have the highest SHAP values for UHPC flowability and compressive strength, respectively. The outcomes of this research study would benefit the scholars of the construction industry to quickly and effectively determine the flowability and compressive strength of UHPC.

Identification of the Elemental Composition of Granulated Blast Furnace Slag by FTIR-Spectroscopy and Chemometrics

Blast furnace slag is a key large-tonnage waste product of metallurgical production, which is considered to be a promising alternative material in construction. In order to determine the scope of potential use of slag as a marketable product, it is necessary to study its structure and composition, which is determined by means of modern analytical instrumental methods. This paper analyzes the application of Fourier transform infrared spectroscopy (FTIR) and chemometrics methods to develop calibration models for identifying pelletized slag by elemental composition. In a comparative analysis of FTIR-spectra of slag the characteristic frequencies of absorption bands responsible for the content of calcite, silicates and aluminosilicates in the composition of samples were determined. Multivariate regression methods (principal components regression, partial least squares regression) and data of elemental composition results by EDX method were used to develop calibration models for determining elemental composition of granulated blast furnace slag. Using the developed PLS models with high performance (R2 from 0.91 to 0.96 for different components), the prediction of the elemental composition (Ca, Si, O, Mg) of the test sample was carried out and a low deviation of the prediction in contrast to the EDX reference data was obtained. The use of PLS calibration models for rapid and nondestructive determination of the quantitative content of components of the composition of granulated blast furnace slag has been proposed.

Flexural Strength Prediction of Steel Fiber-Reinforced Concrete Using Artificial Intelligence

Research has focused on creating new methodologies such as supervised machine learning algorithms that can easily calculate the mechanical properties of fiber-reinforced concrete. This research aims to forecast the flexural strength (FS) of steel fiber-reinforced concrete (SFRC) using computational approaches essential for quick and cost-effective analysis. For this purpose, the SFRC flexural data were collected from literature reviews to create a database. Three ensembled models, i.e., Gradient Boosting (GB), Random Forest (RF), and Extreme Gradient Boosting (XGB) of machine learning techniques, were considered to predict the 28-day flexural strength of steel fiber-reinforced concrete. The efficiency of each method was assessed using the coefficient of determination (R²), statistical evaluation, and k-fold cross-validation. A sensitivity approach was also used to analyze the impact of factors on predicting results. The analysis showed that the GB and RF models performed well, and the XGB approach was in the acceptable range. Gradient Boosting showed the highest precision with an R² of 0.96, compared to Random Forest (RF) and Extreme Gradient Boosting (XGB), which had R² values of 0.94 and 0.86, respectively. Moreover, statistical and k-fold cross-validation studies confirmed that Gradient Boosting was the best performer, followed by Random Forest (RF), based on reduced error levels. The Extreme Gradient Boosting model performance was satisfactory. These ensemble machine learning algorithms can benefit the construction sector by providing fast and better analysis of material properties, especially for fiber-reinforced concrete.