Closed-Form Equation for Estimating Unconfined Compressive Strength of Granite from Three Non-destructive Tests Using Soft Computing Models

Skentou, Athanasia D.; Bardhan, Abidhan; Mamou, Anna; Lemonis, Minas E.; Kumar, Gaurav; Samui, Pijush; Armaghani, Danial Jahed; Asteris, Panagiotis G.

doi:10.1007/s00603-022-03046-9

Cited by 45 publications

(7 citation statements)

References 127 publications

(115 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of optimized ML models was evaluated using multiple statistical metrics such as root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), coefficient of determination R 2 , scatter index (SI), and performance index (PI) (see Table 2). Many researchers utilized these indices to evaluate the predictive performance of different ML models [53][54][55][56][57][58][59][60]. The RMSE measures the average magnitude of the errors between the predicted and actual values, indicating the model's predictive accuracy.…”

Section: Methodsmentioning

confidence: 99%

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Chala

Ray

2023

Applied Sciences

View full text Add to dashboard Cite

Seismic response assessment requires reliable information about subsurface conditions, including soil shear wave velocity (Vs). To properly assess seismic response, engineers need accurate information about Vs, an essential parameter for evaluating the propagation of seismic waves. However, measuring Vs is generally challenging due to the complex and time-consuming nature of field and laboratory tests. This study aims to predict Vs using machine learning (ML) algorithms from cone penetration test (CPT) data. The study utilized four ML algorithms, namely Random Forests (RFs), Support Vector Machine (SVM), Decision Trees (DT), and eXtreme Gradient Boosting (XGBoost), to predict Vs. These ML models were trained on 70% of the datasets, while their efficiency and generalization ability were assessed on the remaining 30%. The hyperparameters for each ML model were fine-tuned through Bayesian optimization with k-fold cross-validation techniques. The performance of each ML model was evaluated using eight different metrics, including root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), coefficient of determination (R2), performance index (PI), scatter index (SI), A10−I, and U95. The results demonstrated that the RF model consistently performed well across all metrics. It achieved high accuracy and the lowest level of errors, indicating superior accuracy and precision in predicting Vs. The SVM and XGBoost models also exhibited strong performance, with slightly higher error metrics compared with the RF model. However, the DT model performed poorly, with higher error rates and uncertainty in predicting Vs. Based on these results, we can conclude that the RF model is highly effective at accurately predicting Vs using CPT data with minimal input features.

show abstract

Section: Methodsmentioning

confidence: 99%

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Chala

Ray

2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…The collected and analyzed data were conducted from previous literature studies. The collected data (total datasets of 420) were divided into three groups datasets; training, testing, and validating 18–53 . The training dataset consisted of 280 data points, while the testing and validating datasets comprised 70 data points 54–56 .…”

Section: Methodsmentioning

confidence: 99%

“…Figure 3 elaborates on the distribution of the CS ranges 73% of the data have clustered in a range of 20–55 MPa, and 23% of the data have a range of 56–120 MPa. Only 4% was recorded at the very early ages of the concrete 18–52 …”

Section: Statistical Evaluationmentioning

confidence: 99%

Analysis and prediction of the effect of Nanosilica on the compressive strength of concrete with different mix proportions and specimen sizes using various numerical approaches

Ali

Muayad

Mohammed

2022

Structural Concrete

Self Cite

View full text Add to dashboard Cite

Introducing nanotechnology in concrete is one of the most significant successes in enhancing the mechanical properties of concrete. It affects the quality of the microstructure of the concrete due to its extremely small nanoparticles leading to a faster reaction due to its large surface area. The nanoparticles fill the concrete's pores, resulting in considerably improved strengths in the early ages of the concrete. This research used six models; linear regression (LR), multilinear regression (MLR), nonlinear regression (NLR), pure quadratic (PQ), interaction (IA), and full quadratic (FQ), to predict the compressive strength of the concrete modified with different Nanosilica contents for various mix proportions. The models were applied to three datasets, with 420 collected from several research studies. The ranges of the concrete mix proportions used in this study were as follows, water/cement ratio (w/c) ranged between 0.1 and 1, cement content (C) ranged between 153.81 and 1200 kg/m3, fine aggregate content (FA) ranged between 492 and 2270 kg/m3, coarse aggregate content (CA) ranged between 617 and 2900 kg/m3, superplasticizer (SP) ranged between 0% and 6.7%, coarse aggregate size (CAS) ranged between 6 and 50.8 mm, fine aggregate size (FAS) ranged between 0.025 and 10 mm, Nanosilica content (NS%) ranged between 0% and 15%, w/c of 0.4–0.6 and the curing time (days) ranged between 3 and 180. The compressive strength of the collected datasets ranged from 3 to 120 MPa. The parameters that have improved the analyzed datasets of the compressive strength were the content of Nanosilica and the SP. The models were assessed for their accuracy using multiple assessment criteria; the most evident ones are; the objective function (OBJ), root mean square error (RMSE), Scatter Index (SI), and the maximum absolute error (MAE). The analytical models and the model comparison operations will be shown in detail in this research, indicating that the most effective and accurate model is the FQ model; where coefficient of determination (R2) of 0.96, RMSE of 3.49 MPa, MAE of 2.96 MPa, and SI of 0.08 were conducted from the analytical studies. From the developed models, the compressive strength of the concrete up to a high compressive strength level with high accurately can be predicted. Also, the compressive strength of the concrete can be accurately predicted for various sizes of aggregate and specimens.

show abstract

“…A. D. Skentou et al, based on artificial neural network, predicted the unconfined compressive strength of granite by using three nondestructive testing indexes: pulse velocity, Schmidt hammer rebound number and effective porosity. Experimental results showed that the performance of the Levenberg-Marquardt artificial neural network designed by the study was superior to the existing prediction model [22]. To solve the design problem of multiple heat recovery technology, M. A. Haghghi et al proposed A new polygeneration model considering parallel and series waste heat recovery, and carried out multi-criteria optimization under four different scenarios by using artificial neural network and multi-objective gray wolf optimization method.…”

Section: Related Workmentioning

confidence: 98%

Clustering Analysis of Unlabeled Data and Weak-Label Detection Analysis Method Integrating Soft Computing Technology

Liang

2024

IEEE Access

View full text Add to dashboard Cite

With the continuous improvement of digitization, the processing and analysis of massive data has become one of the hot issues. Soft computing technology, as an emerging machine intelligence technology, performs well in handling complex uncertainty problems and is an important component of artificial intelligence. This study takes soft computing technology as the technical core and constructs a fuzzy dynamic clustering model based on improved immune algorithms to process unlabeled data. And an anomaly detection and analysis algorithm is designed based on soft instance transfer learning to handle weakly labeled data. The performance test outcomes denote that the accuracy, recall, and F1 values of the immune optimization fuzzy dynamic clustering algorithm are 91.69%, 89.27%, and 92.15%, respectively, reaching the optimal level of similar intelligent optimization clustering algorithms. The immune optimization fuzzy dynamic clustering algorithm has better computational efficiency, loss function curve performance, and strong global search ability, and avoids the occurrence of local optimal solutions. Compared with other advanced clustering algorithms, the immune optimization fuzzy dynamic clustering algorithm performs well on datasets in various fields, and both external and internal evaluation indicators verify the algorithm's clustering effect. The AUC value of the soft computing instance transfer learning anomaly detection algorithm is 0.913, with a detection accuracy of 91.67%, which is superior to other anomaly detection algorithms. The unlabeled and weak-label data processing model designed based on soft computing technology can effectively achieve the processing and analysis of real-world data problems.

show abstract

Closed-Form Equation for Estimating Unconfined Compressive Strength of Granite from Three Non-destructive Tests Using Soft Computing Models

Cited by 45 publications

References 127 publications

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Machine Learning Techniques for Soil Characterization Using Cone Penetration Test Data

Analysis and prediction of the effect of Nanosilica on the compressive strength of concrete with different mix proportions and specimen sizes using various numerical approaches

Clustering Analysis of Unlabeled Data and Weak-Label Detection Analysis Method Integrating Soft Computing Technology

Contact Info

Product

Resources

About