Prediction of Pullout Behavior of Belled Piles through Various Machine Learning Modelling Techniques

Area of Science:

• Geotechnical Engineering
• Computational Intelligence
• Machine Learning Applications

Background:

• Accurate estimation of pullout forces is crucial for the design and stability of belled piles.
• Traditional methods may not fully capture the complex soil-structure interactions involved in pullout behavior.
• Developing advanced modeling techniques can improve prediction accuracy and engineering design.

Purpose of the Study:

• To develop and compare various modeling techniques for estimating the pullout forces of belled piles.
• To evaluate the performance of feedforward neural network (FFNN), radial basis functions neural networks (RBNN), general regression neural network (GRNN), and adaptive neuro-fuzzy inference system (ANFIS).
• To determine the most reliable and accurate model for predicting the pullout behavior of belled piles.

Main Methods:

• Utilized a hybrid learning algorithm (back-propagation and least square estimation) to train the ANFIS model in MATLAB.
• Developed and trained FFNN, RBNN, GRNN, and ANFIS models using 432 samples (300 for training, 132 for testing).
• Evaluated model performance using statistical indexes: coefficient of determination (R), variance account for (VAF), and root mean square error (RMSE).

Main Results:

• ANFIS achieved the highest accuracy for training datasets with R=0.998, VAF=97.442, and RMSE=0.058.
• For testing datasets, ANFIS also demonstrated superior performance with R=0.995, VAF=96.247, and RMSE=1.252.
• Compared to FFNN, RBNN, and GRNN, ANFIS showed significantly better reliability in estimating pullout forces.

Conclusions:

• The adaptive neuro-fuzzy inference system (ANFIS) is a highly reliable and accurate tool for estimating the pullout behavior of belled piles.
• ANFIS outperforms traditional neural network models like FFNN, RBNN, and GRNN in predicting pullout forces.
• The developed ANFIS model provides a robust solution for geotechnical engineers in assessing the capacity of belled piles.