Prediction of Concrete's Compressive Strength via Artificial Neural Network Trained on Synthetic Data

Alghamdi, Saleh J.

doi:10.48084/etasr.6560

Cited by 4 publications

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the field of material modeling, some researchers [9][10][11] have used a back-propagation ANN to model the behavior of concrete in the state of plane stress under monotonic biaxial loading and compressive uniaxial cycle loading. Their findings appear to be very promising.…”

Section: Introductionmentioning

confidence: 99%

Pile Design using the Modified Unified Method combined with Monte Carlo Simulation

Cao Van

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

Piles are typically designed to ensure the bearing capacity and settlement elastic behavior. However, some projects seem over-designed, leading to unnecessary waste, whereas others experience excessive settlement. This could be caused by various factors, such as site investigation, sampling and testing methods, selection of soil behavior model, and calculation programs. To achieve a successful pile design, engineers must consider, among others, the loads applied to the pile, the resistance capacity of the piles, the pile material's bearing capacity, the pile's displacement, and the soil's settlement. On the other hand, the input parameters for geotechnical problems, in general, and pile design problems, in particular, often do not reflect the actual behavior of the soil due to its heterogeneous and anisotropic nature. To address these challenges, an Artificial Neural Network (ANN) approach is proposed for pile design, using a relatively wide range of soil input data. This study establishes a numerical program for pile design combined with the ANN approach, validated by verifying the pile design of a project constructed in Vietnam. The results indicate that the proposed program can reasonably simulate pile group behavior and assist engineers in deploying appropriate safety factors.

show abstract

Section: Introductionmentioning

confidence: 99%

Pile Design using the Modified Unified Method combined with Monte Carlo Simulation

Cao Van

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

show abstract

“…The design parameters of the equivalent block were recommended many years ago [1,2]. The design parameters of the equivalent stress block have been verified in many times [3][4][5][6][7]. Authors in [3] developed an equivalent stress block for high strength concrete in revision of the Indian Code, following the same procedure followed in [1].…”

Section: Introductionmentioning

confidence: 99%

Modified Equivalent Compression Stress Block for Normal-Strength Concrete Flexural Design using Energy Modeling

El-Gohary

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

The equivalent stress block is recommended for use in the design of reinforced concrete sections to simplify the analysis of the composite behavior of concrete and steel reinforcement. In most current codes, a rectangular equivalent stress block is provided. The design parameters of the equivalent block were recommended many years ago. Due to the importance of the equivalent stress block concept, numerous investigations have been performed to increase its accuracy. In the current paper, an exploration of the rectangular equivalent stress block has been carried out using the energy modeling approach. Energy modeling is a new general approach for studying the behavior of concrete elements. In this method, the energy consumed (work done) can be determined by integrating the force-displacement diagram (in the current study this will be the concrete stress-strain curve in compression). Schematic and equivalent stress-strain curves for concrete in uniaxial compression provided in most current codes and relevant textbooks were considered in this research. The codes taken into account in the current study are ACI-318-19, Canadian Code CSA A23.3-04, Eurocode EC-2, and Chinese standard GB 500 10 – 2002. The energy consumed by these curves for different values of concrete strength has been compared with numerous experimental results. This comparison shows that the results of the equivalent stress block provided in most of the considered current codes are conservative. Applying the energy modeling for the considered experimental stress-strain curves a modified equivalent stress block is recommended for practical use. The results of the proposed equivalent stress block are in good agreement with the experimental ones. The ratio between the predicted total energy engaging the proposed model and the total energy calculated for the experimental results ranges between 0.95 and 1.08 with a mean value equal to unity.

show abstract

“…Therefore, techniques such as data augmentation (Shorten and Khoshgoftaar 2019; Mumuni and Mumuni 2022) and synthetic data generation (Hong et al 2021) have been used to improve data quality and to obtain adequate and representative training data for ML when data on the subject are limited or do not have good quality. Synthetic data are statistically generated from real sample data to increase the volume of ML model training and development data (Wendland et al 2022) and have been used to generate data in many elds of knowledge, such as chemistry, medicine, nance, marketing and engineering, when the availability of real data is limited (Alghamdi, 2023;Gelernter et al, 1990;Hittmeir et al, 2019;Krüger Mariusand Vogel-Heuser, 2024).…”

Section: Introductionmentioning

confidence: 99%

On the use of Synthetic Data for Machine Learning prediction of Self-Healing Capacity of Concrete

de Oliveira,

Stefani

2024

Preprint

View full text Add to dashboard Cite

This work investigated the use of synthetic data to overcome the limitations of scarce experimental data in predicting the self-healing capacity of bacteria-driven concrete. We generated a synthetic dataset based on real-world data, significantly expanding the original dataset and then trained and compared machine learning models, including probabilistic and ensemble methods, to predict the concrete self-healing capacity. The results demonstrate that the ensemble methods, particularly the random forest (RF) method (accuracy = 0.863 and F1-score = 0.863), outperformed the probabilistic models and achieved high accuracy in predicting self-healing capacity. The trained models were further applied to real-word data examples, showing high accuracy. This research validates the utility of synthetic data in predicting modelling accuracy and reliability in civil engineering, particularly in areas with limited experimental data. The findings contribute to the growing use of ML and AI in concrete research and demonstrate the transformative potential of synthetic data in addressing challenges in civil engineering.

show abstract

Prediction of Concrete's Compressive Strength via Artificial Neural Network Trained on Synthetic Data

Cited by 4 publications

References 24 publications

Pile Design using the Modified Unified Method combined with Monte Carlo Simulation

Pile Design using the Modified Unified Method combined with Monte Carlo Simulation

Modified Equivalent Compression Stress Block for Normal-Strength Concrete Flexural Design using Energy Modeling

On the use of Synthetic Data for Machine Learning prediction of Self-Healing Capacity of Concrete

Contact Info

Product

Resources

About