Análise geoestatística da estimativa de resistência estátistica a partir de ensaios SPT

Abstract

Prediction of the soil-structure static resistance is primordial in foundation engineering design projects. Use of methods for prediction of pile bearing capacity based on standard penetration test (SPT) results is a common practice. Simplicity and low costs make the SPT one of the most employed tests in geotechnical investigation. In the last four decades, the mechanisms of energy transfer within the test have been studied by several authors, who were motivated by questions related to the dispersion of the blow-count values. It has been verified that knowing the test energy efficiency allows to normalize the SPT blow-count resistance to a standard reference value. In this context, Aoki and Cintra (2000) investigated the dynamic SPT hammer blow event by means of the Hamilton’s energy conservation principle. The study was a theoretical background to the methodology put forward by Aoki (2013), which allows to quantify pile tip and side friction resistances from the SPT test results. The method can be applied to tests in which the efficiency and the length of the recovered of soil samples are known. In the present study, predictions of pile tip and side friction static resistances were achieved by using the methodology of Aoki (2013) and the results of 58 SPT tests. Borings were carried out in a site composed of fine sand layers. All tests had the length of the recovered samples measured. An additional SPT test was also carried out, which included six static pullout tests on the SPT sampler, in order to investigate the lateral skin friction of the sampler. The ratio between the internal and the external skin friction of the sampler, known as the “a” parameter, was close to two for fine sand and increased for deeper layers with higher percentage of fines. Finally, a comparison was made between the results of resistance predicted by Aoki (2013) method and the results of eight CPT tests performed at the same site. Comparisons between predicted and experimental results were based on concepts of Geostatistics. Specifically, an analysis of the spatial behavior of the static resistance was conducted using the software Geokrige, which made calculations of variograms and spatial inferences by ordinary kriging. Although the magnitude of the side friction static resistance values predicted by the method proposed by Aoki (2013) corresponded to that of the field tests, a large dispersion in the values has been identified. However, most predictions of pile toe resistance were successful. Ratios of predicted to experimental toe resistance values ranged from 1.0 to 1.6 and remained within acceptable dispersion levels. Predictions of pile toe static resistances based on SPT tests were found to be consistent.