Incorporation of artificial neural network with principal component analysis and cross-validation technique to predict high-performance concrete compressive strength
Compressive strength is the most essential mechanical characterization for concrete due to its crucial role in stating the design standards. Therefore, early, and accurate evaluation of concrete compressive strength minimizes efforts, costs, and time. In this study, we investigate the ability of art...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Springer
2021
|
| Subjects: | |
| Online Access: | http://eprints.um.edu.my/35743/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Compressive strength is the most essential mechanical characterization for concrete due to its crucial role in stating the design standards. Therefore, early, and accurate evaluation of concrete compressive strength minimizes efforts, costs, and time. In this study, we investigate the ability of artificial neural network (ANN) incorporated with principal component analyses (PCA) and cross-validation (CV) techniques to forecast the high-performance concrete (HPC) compression strength. The obtained results from the ANN-CVPCA model showed a good agreement between predicted and actual values. The proposed model provides high accuracy prediction of HPC compressive strength. It also provided a higher correlation coefficient (0.96) and a lower value of mean absolute error (3.43mpa), root mean square error (4.64mpa) and normalized root mean square error (0.13). Moreover, a sensitivity analysis was carried out to identify the most influential parameters and the simulated results showed that the superplasticizer, blast furnace slag, and cement parameters respectively have great effects on the compressive strength of HPC. The performance of the ANN-CVPCA model compared with other models published in previous studies and achieved the desired superiority and more stable predictions due to the existence of PCA and CV which play a significant role in increasing the generalization ability as well as avoiding redundant data and reducing the uncertainty in modeling outcomes. © 2021, The Author(s), under exclusive licence to Springer Nature Switzerland AG. |
|---|