Forecasting the Bearing Capacity of Open-Ended Pipe Piles Using Machine Learning Ensemble Methods

The bearing capacity of open-ended pipe piles is an important geotechnical engineering problem with significant practical implications in the construction industry. In recent years, basic machine learning methods have gained popularity for their ability to predict the bearing capacity of such piles accurately. In this paper, we propose an ensemble machine-learning approach to predict the bearing capacity of open-ended pipe piles. We compare the performance of several popular ensemble methods, including bagging, boosting, stacking, and voting; and assess the accuracy of our proposed approach using real-world data. Our results show that the ensemble approach outperforms individual machine learning models, yielding more accurate predictions of the bearing capacity of open-ended pipe piles. The proposed approach can potentially be applied in practice to improve the design and construction of open-ended pipe pile foundations. In addition, the proposed approach exhibited satisfactory diameter and length effects, which have been areas of concern for some traditional design approaches. The work thus demonstrates the feasibility of employing machine learning (ML) for determining the capacity of pipe piles.