A quantitative approach to estimating pile bearing capacity using multidimensional datasets and novel modeling techniques

The project discusses a modern approach to estimating pile-bearing capacity (PBC), a critical subject in geotechnical engineering that affects the construction, design, and safety of a foundation. Particularly, PBC addresses the amount of load piling that can take and sustained without a risk of exc...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Multiscale and Multidisciplinary Modeling, Experiments and Design Jg. 9; H. 1; S. 28
Hauptverfasser: Li, Huijing, Yang, Zhangli
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Cham Springer International Publishing 01.12.2026
Springer Nature B.V
Schlagworte:
Accuracy
Algorithms
Artificial intelligence
Bridge failure
Characterization and Evaluation of Materials
Decision trees
Engineering
Estimation
Expected values
Foundation failure
Geotechnical engineering
High rise buildings
Load
Machine learning
Mathematical Applications in the Physical Sciences
Mechanical Engineering
Modelling
Numerical and Computational Physics
Offshore structures
Optimization
Original Paper
Pile bearing capacities
Preliminary designs
Safety engineering
Simulation
Solid Mechanics
Predictive modeling
Geotechnical engineering
Pile bearing capacity
Hybrid schemes
Machine learning
ISSN:2520-8160, 2520-8179
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The project discusses a modern approach to estimating pile-bearing capacity (PBC), a critical subject in geotechnical engineering that affects the construction, design, and safety of a foundation. Particularly, PBC addresses the amount of load piling that can take and sustained without a risk of excessive settlement or failure of the structure, including high-rise buildings, bridges, and offshore structures. Estimating this capacity is significant to the foundation. Its underestimation leads to foundation failure, while its overestimation leads to increased costs through material procurement or deeper foundations than are needed. In this work, two machine learning (ML) heuristics of decision tree regression (DTR) and voting regression (VR) have been applied to boost the prediction accuracy related to PBC, since traditional empirical and analytical methods generally do not provide adequate protection to the interaction complexities in soil-pile systems. Two innovative optimization algorithms have been implemented for the optimization of the modeling technique these are the arithmetic optimization algorithm (AOA) and the equilibrium slime mold algorithm (ESMA). The performances of the techniques were reviewed using five evaluation metrics, R² and RMSE (kN) being among them. Outcomes showed that the best model was the VR optimized with ESMA, VRES, which had an R² of 0.979 and an RMSE of 224.665. Hence, the scheme performed best in this study. These predictive schemes optimize engineering decision-making processes and thus can be used in preliminary design and safety assessment phases of foundation projects. Proper estimation of pile-bearing capacity by the engineer can minimize construction risks, optimize material use, and improve efficiency.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2520-8160
2520-8179
DOI:10.1007/s41939-025-01093-7