Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Structural Engineering and Mechanics Volume 91, Number 6, September25 2024 , pages 583-589 DOI: https://doi.org/10.12989/sem.2024.91.6.583 |
|
|
|
Study on the applicability of regression models and machine learning models for predicting concrete compressive strength |
||
Sangwoo Kim, Jinsup Kim, Jaeho Shin and Youngsoon Kim
|
||
| Abstract | ||
| Accurately predicting the strength of concrete is vital for ensuring the safety and durability of structures, thereby contributing to time and cost savings throughout the design and construction phases. The compressive strength of concrete is determined by various material factors, including the type of cement, composition ratios of concrete mixtures, curing time, and environmental conditions. While mix design establishes the proportions of each material for concrete, predicting strength before experimental measurement remains a challenging task. In this study, Abrams's law was chosen as a representative investigative approach to estimating concrete compressive strength. Abrams asserted that concrete compressive strength depends solely on the water-cement ratio and proposed a logarithmic linear relationship. However, Abrams's law is only applicable to concrete using cement as the sole binding material and may not be suitable for modern concrete mixtures. Therefore, this research aims to predict concrete compressive strength by applying various conventional regression analyses and machine learning methods. Six models were selected based on performance experiment data collected from various literature sources on different concrete mixtures. The models were assessed using Root Mean Squared Error (RMSE) and coefficient of determination (R2) to identify the optimal model. | ||
| Key Words | ||
| Abrams's law; compressive strength; concrete; machine learning; prediction model | ||
| Address | ||
| Sangwoo Kim, Jinsup Kim: Department of Civil Engineering, Gyeongsang National University, Jinju 52828, Republic of Korea Jaeho Shin: Department of Information and Statistics, Gyeongsang National University, Jinju 52828, Republic of Korea Youngsoon Kim: Department of Information and Statistics and Dept. of Bio & Medical Bigdata (BK21 Four Program), Gyeongsang National University, Jinju 52828, Republic of Korea | ||