The Probability Design of Pile Foundation (Design Pondasi Tiang dengan Prinsip Probabilitas)

Abstract

Probabilistic design began more realistic solution compare to the conventional deterministic design as it can describe and accommodate of soil variability. Although the old approach is widely used, it is confirmed, that almost all natural soils are spatially variable in their properties and rarely homogenous. This paper focuses on the analysis and design of pile foundation on the project of SMK Muhammadiyah Sukoharjo based on the probabilistic theory. The research begins with a certain preliminery analysis to prepare the probabilistic analysis of Pile Foundation design by characterizing the tip resistance on 3 CPTs data taken from the site. It involves an extensive analysis to perform the best-fit distribution of pointwise variability of tip resistance using computer program written in MATLAB. The next step is deriving the point statistics (i.e. mean, standard deviation, and coefficient of variation) across the site. Finally, a certain monte carlo analysis using CRYSTAL BALL computer program is carried out to design the pile foundation by inputting all data taken from MATLAB analysis. The pile is installed in 10m depth with the thickness of cap and the diameters of pile are 0.4m and 0.3m respectively. The pile experiences a 140 tones vertical load.The results show that, there is no objection to the hypothesis of normality in the chisquare analysis, although the best fit distribution for each profile or 6 profiles which collected at once are varying (i.e.normal, log-normal, gamma). Compare to the deterministic analysis, the probabilistic design needs more pile to resist the vertical load in a safety factor of 2.5 and confident limit 95%. Furthermore, the higher variation of CPT data the larger number of pile is desired.