Estimating the Pile Settlement Using a Machine Learning Technique Optimized by Henry's Gas Solubility Optimization and Particle Swarm Optimization
Ensuring constructional projects are safe, like stacked structures, requires consideration to immunize structures over the period. Pile settlement (PS) is an important project problem and is receiving a lot of attention to prevent failure before construction starts. Several items for estimating pile...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Bilijipub publisher
2022-12-01
|
| Series: | Advances in Engineering and Intelligence Systems |
| Subjects: | |
| Online Access: | https://aeis.bilijipub.com/article_163964_79fbf8ec9816c1ae968f8abc638e8eb3.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Ensuring constructional projects are safe, like stacked structures, requires consideration to immunize structures over the period. Pile settlement (PS) is an important project problem and is receiving a lot of attention to prevent failure before construction starts. Several items for estimating pile motion can help understand the project's perspective during the loading phase. Most intelligent strategies for the mathematical calculation of pile movement are used in PS simulations. Therefore, in present article, a developed framework operating support vector regression (SVR) together with Henry's Gas Solubility Optimization (HGSO) and Particle Swarm Optimization (PSO) was considered for accurate pile motion calculation. The usages of optimizers were to tune some internal settings of SVR. The Kuala Lumpur transportation network was selected to study the movement of piles based on the land rock characteristics using the developed SVR-HGSO and SVR-PSO structures. Five metrics were used to evaluate the performance of each model. The main objective of this research is to evaluate the artificial inteligent approach in form of two developed models in simulating the pile settlement rates using hybrid optimized frameworks. The R2 of modeling both were obtained similarly at 0.99 level. While the RMSE of SVR-PSO appeared more than two-fold of SVR-HGSO, 0.46 and 0.29 mm, respectively. Also, test phase results showed the better performance of SVR-HGSO with an MAE index of 0.278, which is 57.10% lower than the other one. The OBJ proved accurate modeling by SVR-HGSO calculated at 0.283mm level. |
|---|---|
| ISSN: | 2821-0263 |