- author = "Yaser Moodi and Naser Safaeian Hamzehkolaei and Iman Afshoon",
- title = "Intelligent Models for Predicting the Compressive Strength of Green Concrete Made with Fine and Coarse Grains of Waste Copper Slag",
- journal = "Karafan Journal",
- year = "2024",
- volume = "21",
- number = "3",
- pages = "367--394",
- keywords = "genetic algorithms, genetic programming, PSO, Green concrete, copper slag, Concrete Aggregates Compressive Strength Prediction Models, ANFIS",
- publisher = "National University of Skill",
- ISSN = "2382-9796",
-
URL = "
https://karafan.tvu.ac.ir/article_212355.html?lang=en",
-
URL = "https://karafan.tvu.ac.ir/article_212355.html",
-
URL = "
https://karafan.tvu.ac.ir/article_212355_7637d88f811623fa8e61f99f07ab8447.pdf",
-
DOI = "
10.48301/kssa.2024.441580.2833",
- size = "28 pages",
- abstract = "In recent years, using copper slag waste as a substitute for aggregate materials in concrete has become popular due to economic savings and environmental benefits. However, the time-consuming and costly nature of the experimental studies has prompted the use of artificial intelligence-based models to predict concrete properties. Thus far, no method has been proposed to estimate the compressive strength of concretes with copper slag as a substitute for natural aggregates. This study presents an effective relationship for estimating the compressive strength of such concrete based on a robust database of 458 samples from valid laboratory studies. The dataset included variables such as the ratio of water-to-powdered materials, amount of powdered materials, amount of fine and coarse aggregates, copper slag percentage, and concrete curing age. Two intelligent models, Genetic Programming (GP) and Adaptive Neuro-Fuzzy Inference System (ANFIS), were used. The Particle Swarm Optimization (PSO) algorithm was employed to tune parameters and optimize the ANFIS model. Results showed that the ANFIS model (R2=0.94) outperformed the GP model in generalization capability and accuracy. The hybrid ANFIS-PSO model with optimal parameter tuning achieved the best performance (R2=0.96) compared to other models. Through sensitivity analysis of the input variables, it was determined that curing age and powdered materials had the highest positive effects, respectively, while the ratio of water-to-powdered material had the most negative effect on the compressive strength of green concrete containing copper slag. An increase in the volume of fine aggregates also led to a significant decrease in the compressive strength of this type of concrete. The proposed GP-based predictive model enables the estimation of compressive strength and the conduct of supplementary parametric studies without the need for complex calculations and additional costs.",
-
notes = "In Persian with some English translation
Department of Civil Engineering, Sirjan University of Technology, Sirjan, Iran",
