Feature engineering for predicting compressive strength of high-strength concrete with machine learning models

Kumar, Praveen; Pratap, Bheem

doi:10.1007/s42107-023-00807-x

Cited by 23 publications

(3 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If the AIV greater than 30 %, the results should be treated with caution [16]. With the application of machine learning models, the results obtained from this experiment can be numerically predicted as stated in [17] which was used to predict the compressive strength of high-strength concrete.…”

Section: Aggregate Impact Value (Aiv)mentioning

confidence: 99%

Recycled concrete aggregate and cinder gravel as base course construction materials

Sahile,

Agon,

Geremew

et al. 2024

ACAE

View full text Add to dashboard Cite

In this research, the potential use of recycled concrete aggregate (RCA) and CG as a base course construction material was investigated. A laboratory test involving the mechanical stabilization of RCA and CG was conducted to examine their physical properties. Eight samples of RCA blended with CG in varying proportions from 0% to 100 % with 10 % variation were studied. The laboratory test results indicate that 100 % CG yields specific gravity (SG), aggregate crushing value (ACV), aggregate impact value (AIV), Los Angeles abrasion (LAA), flakiness index (FI), elongation index (EI), plasticity index (PI), water absorption, soundness, and California bearing ratio (CBR) values of 2,54 %, 38,37 %, 20,10 %, 33,17 %, 5,59 %, 12,09 %, 0,90 %, 3,52 %, 10,60 %, and 38,08 %, respectively. The results for 100% RCA show SG, ACV, AIV, LAA, FI, EI, PI, water absorption, soundness, and CBR values of 2,70 %, 9,56 %, 5,30 %, 9,20 %, 15,30 %, 15,84 %, NP, 0,23 %, 1,49 %, and 105,87 %, respectively. These results also fail to meet the gradation requirements based on ERA standard specifications. Therefore, mechanical stabilization was adopted to improve the physical properties of the samples. Blending 60 % RCA with 40 % CG resulted in SG, ACV, AIV, LAA, FI, EI, PI, water absorption, soundness, and CBR values of 2,65 %, 19,46 %, 10,70 %, 10,88 %, 16,15 %, 22,01 %, NP, 0,26 %, 2,09 %, and 101,98 %, respectively. At this proportion, the gradation aligns with the required ERA standard specifications for GB2 and GB3 materials. Therefore, CG up to 40 % by weight with 60 % RCA is viable for road base course construction, especially when readily available or nearby.

show abstract

Section: Aggregate Impact Value (Aiv)mentioning

confidence: 99%

Recycled concrete aggregate and cinder gravel as base course construction materials

Sahile,

Agon,

Geremew

et al. 2024

ACAE

View full text Add to dashboard Cite

show abstract

“…It is utilised to develop and validate ML models in this context. Specifically, an artificial neural network (ANN) [45,49] was utilised for this study, integrated within the Python programming interface. The data was normalised and scaled using the standard scalar technique.…”

Section: Figure 2 Scanning Electron Microscopy (Sem) Of Eggshell At D...mentioning

confidence: 99%

“…These models mainly employ regression-based methods to estimate construction material properties. Notably, machine learning (ML) techniques [45,46] have emerged as key contributors to the This study focuses on predicting concrete strength sequentially. It encompasses descriptions of the materials and methodology employed and the generation of dataset points through experimental programs.…”

Section: Introductionmentioning

confidence: 99%

Soft computing techniques for analysing the mechanical properties of egg shell powder-based concrete

Sharma,

Kumar,

Bano

et al. 2024

ACAE

View full text Add to dashboard Cite

The construction industry is increasingly focused on sustainability to reduce environmental impact. Researchers are actively exploring alternative materials to replace clinker-based binders. This study specifically investigates the use of eggshell powder (ESP) as a sustainable substitute in construction. Portland slag cement (PSC) is partially replaced by ESP in concrete production for this purpose. To assess the effectiveness of ESP in enhancing binder properties, the study analyses experimental data for compressive and flexural strength. Artificial Neural Network (ANN) modelling is employed for this analysis to predict material performance. The model undergoes training and testing using input data to ensure accuracy and reliability. The success of the study is demonstrated by high R2 values, with 0,9915 for compressive strength and 0,9921 for flexural strength, indicating that the ANN model closely matches actual material performance. Additionally, error analysis confirms the model's remarkable accuracy in predicting real-world results. Furthermore, the research highlights the exceptional potential of the developed ANN model, which can effectively predict the mechanical properties of construction materials containing ESP.

show abstract

Review of Recent Advances on AI Applications in Civil Engineering

Aydın,

Bekdaş,

Nigdeli

2024

Springer Tracts in Nature-Inspired Computing

View full text Add to dashboard Cite

Feature engineering for predicting compressive strength of high-strength concrete with machine learning models

Cited by 23 publications

References 33 publications

Recycled concrete aggregate and cinder gravel as base course construction materials

Recycled concrete aggregate and cinder gravel as base course construction materials

Soft computing techniques for analysing the mechanical properties of egg shell powder-based concrete

Review of Recent Advances on AI Applications in Civil Engineering

Contact Info

Product

Resources

About