Abstract:
This thesis presents an evaluation of the performance of eight methods based on cone
penetration test (CPT) for predicting the ultimate load carrying capacity of square precast RC
concrete piles at a site in Siddhirganj. The following methods were used to predict the load
carrying capacity of the piles using the CPT data: Schmertmann, Bustamante and Gianeselli
(LCPC/LCP), de Ruiter and Beringen, Tumay and Fakhroo, Price and Wardle, Philipponnat,
Aoki and De Alencar, and Penpile method. The ultimate load carrying capacity for each pile
is also predicted using the traditional method based on SPT. The ultimate pile capacity
obtained from CPT data is also compared graphically with the traditional method based on
SPT data.
Prediction of pile capacity has been performed at seven locations within the site. However,
Evaluation of the prediction methods was conducted using the statistical analysis based on
the results of six friction piles and one end-bearing pile.
An evaluation scheme has been executed to evaluate the CPT methods based on their ability
to predict the ultimate pile capacity. Only the criteria selected to evaluate the performance of
the prediction methods is: the arithmetic mean and the standard deviation. Based on this
evaluation, the Schmertmann method, de Ruiter and Beringen method and Philipponnat
method show the best performance in predicting the load carrying capacity of square precast
RC concrete piles. But Philipponnat method does not consider the consistency of cohesive
soil (soft or hard) to determine the ultimate shaft friction capacity,Qs of pile and assumes the
same empirical factor for all types of clay. Aoki and De Alencar method exhibits moderate
performance and Bustamante and Gianeselli (LCPC/LCP) method shows unsatisfactory
performance to estimate the pile capacity. The worst prediction methods are Penpile method
and Price and Wardle method, which are very conservative (underpredict the pile capacities)
and the Tumay and Fakhroo method, which overpredicts the pile capacity excessively. The
four CPT methods, which are de Ruiter and Beringen method, Philipponnat method,
Schmertmann method and Aoki and De Alencar method show better performance than the
currently used method based on SPT. The soil of this site is very erratic and the thickness of
soil layers varies drastically throughout the site. SPT is not reliable for cohesive soil which is
the predominant soil of the whole site.
