A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
The settlement of piles socketed into rock has captured remarkable consideration. In spite of very design techniques of settlement of pile are introduced, developing new as well statistical method with higher performance prediction may be of advantage. At the present study, multivariate adaptive regression splines (MARS) method with different degree of interactions were proposed to predict the settlement of piles in the Klang Valley Mass Rapid Transit (KVMRT) project that constructed in Kuala Lumpur, Malaysia. In this research the results of pile driving analyzer tests along with piles and earth’s properties were considered. Five parameters considered as inputs and piles settlement as output variable. Comparing the measured settlements from experimental efforts with those predicted by MARS method are supplied that the developed models have coefficient of determination (R2) in acceptance rate. By apprising the performance index (PI) index as the whole model evaluator, which considers other indexes altogether, the order 3 MARS model outperforms the other two models, with lower PI values equal to 0.0541 and 0.0951 in training and testing phase, respectively. Therefore, the third interaction equation of MARS for predicting settlement of rock-socketed piles can be recognized as the proposed regression model.
The settlement of piles socketed into rock has captured remarkable consideration. In spite of very design techniques of settlement of pile are introduced, developing new as well statistical method with higher performance prediction may be of advantage. At the present study, multivariate adaptive regression splines (MARS) method with different degree of interactions were proposed to predict the settlement of piles in the Klang Valley Mass Rapid Transit (KVMRT) project that constructed in Kuala Lumpur, Malaysia. In this research the results of pile driving analyzer tests along with piles and earth’s properties were considered. Five parameters considered as inputs and piles settlement as output variable. Comparing the measured settlements from experimental efforts with those predicted by MARS method are supplied that the developed models have coefficient of determination (R2) in acceptance rate. By apprising the performance index (PI) index as the whole model evaluator, which considers other indexes altogether, the order 3 MARS model outperforms the other two models, with lower PI values equal to 0.0541 and 0.0951 in training and testing phase, respectively. Therefore, the third interaction equation of MARS for predicting settlement of rock-socketed piles can be recognized as the proposed regression model.
Settlement prediction of the piles socketed into rock using multivariate adaptive regression splines
QiLi Zuo (author)
2022
Article (Journal)
Electronic Resource
Unknown
Metadata by DOAJ is licensed under CC BY-SA 1.0
Case Studies of Rock-Socketed Piles
| British Library Online Contents | 1996
Case Studies of Rock-Socketed Piles
| Online Contents | 1996
Elastic Analysis of Rock-Socketed Piles
| ASCE | 2015
ELASTIC ANALYSIS OF ROCK-SOCKETED PILES
| British Library Conference Proceedings | 2015
A simplified model for single pile settlement calculation of rock-socketed piles
| Tema Archive | 1999