Numerical simulation of bi-directional osterberg load tests performed in limestone by using finite element analysis

Abstract

Pile load tests are one of the most important geotechnical eld tests. Piles are loaded in practise by the structure from the top. However especially for piles with large load capacities, it is very di cult to simulate this loading method. To overcome this di culty, bi-directional pile load test which is also called Osterberg Cell load test has been developed. Within the scope of this thesis, the theory which is used for converting load-settlement behaviour obtained during bi-directional pile load test into equivalent load-settlement behaviour in ideal top-down load test has been veri ed. To compare the pile behaviour during these two tests; test results of three large-diameter instrumented piles socketed in limestone and performed in Moscow are back-analysed by using nite element analysis. Until load-displacement behaviour obtained during the eld test is simulated and unit shaft frictions measured via strain gauges are obtained, soil parameters of analysis model are calibrated and ideal top loaded pile behaviour generated by these parameters is investigated. As a result of the analyses, it is seen that if the strain gauges are not placed in the correct position, incorrect data can be measured during pile load test and participating this data in the evaluation could lead to miscalculation of the unit shaft friction and consequently the end bearing resistance. Comparing O-cell load test with the ideal top loading test, it is observed that the unit shaft friction values are the same but the locations where they are observed are di erent. It is also observed that the pile bearing capacity and pile behaviour could not be predicted correctly when the applied test load is much smaller than the pile bearing capacity and thus the settlements measured during the test are very small.