High quality data on axial pile load tests in clay were selected from the public domain and put into a data base. Subsequently, statistical analyses of these data were performed to first assess the reliability of current offshore design criteria and, secondly, to develop a statistically more reliable design method. These analyses reveal that the ratio between average ultimate skin friction and undrained shear strength (=alpha) reduces with increasing degree of overconsolidation and increasing pile slenderness (=Pile length over diameter ratio). Soil plasticity is shown to have a negligible influence on this correlation. A simple algorithm accounting for these parameters is obtained from statistical analyses. Predictions of pile capacity made using this new method are shown to be more accurate than those made using conventional current design criteria.
Various reviews of axial pile load test data on pipe piles driven into clays were performed in the mid eighties The result of these reviews are reflected in the foundation design criteria presented in API RP2A4'5. Since then, various new high quality pile loal test data have become available. These tests often showed signification different capacities than predicted using the API W2A criteria, In view of these discrepancies the authors undertook a thorough study of pile load test data. The study consisted of review and selection of high quality pipe load test data, followed by a detailed comparison of these data with pile capacities predicted using API RP2A. This study revealed shortcomings in these criteria. Subsequently a new design method was developed which largely alleviated these shortcomings.
This next study was done in two phases, The first phase (development) consisted of developing the most appropriate formulation of a new design method, This made use of a large number of statistical analyses of test data. Each pile load test was characterized by an average friction value and a set of average soil parameters along the entire pile length. The second phase (verification) consisted of checking the developed design method. In this phase detailed soil profiles for pile load test sites were put in a large data base programme, where after pile capacities were predicted using the newly developed criteria. These predictions were compared with the observed ultimate pile capacities. The results confirm that the new method is signification more reliable than API WA for predicting static pile capacity.
Twenty-six pile load tests were selected from the public literature for this study '22 The selection of high quality ‘old’ (i,e. pre-1984) pile bad test data was based on earlier research work by Sample and Rigdens and by 01son3. Sample and Rigdens developed a set of pile load test data from Olson's data base, In this data set several pile load test data at one site were averaged. This was done to avoid clutter of data points and giving undue weight to essentially repetitive result one soil-pile combination.
