An Appraisal of API RP2A Recommendations for Determining Skin Friction of Piles in Sand

Abstract

ABSTRACT API RP2A in its 14th edition (1984) recommends the use of an earth pressure coefficient of 0.8 for computing the tensile skin friction of an open ended pile driven into silica sands. In view of the impact of this recommendation on the design of North Sea piles driven into sands, a study was made of an expanded data base on the pull-out capacity of piles driven into sands. The expanded data base includes test data on relatively long North Sea conductors and in all 24 tension tests considered of reasonablyhigh quality were analysed. The new data base together with observations of pile behavior in sand has enabled the authors to develop two procedures for designing piles in sand. These two procedures differ from API in allowing skin friction to reduce at a given level as a pile is driven deeper. Comparisons of predicted and measured capacities of the test piles were made and compared to API (1984) predictions. It is concluded that the API Method is not reliable for loose or very dense sand sites. The two new methods provide much better predictions in these soil conditions. INTRODUCTION The Document API RP2A (Ref. 13) is updated regularly in the light of experience and research. From the foundations point of view, a major change was introduced in the 1984 issue (14th Edition) with regard to the assessment of skin friction of pilesdriven into sands. The 1984 and subsequent issues of API recommend using a coefficient of earth pressure, K, value of 0.8 for computing both compressive and tensile friction capacities of an open ended pile driven into silica sands. This recommendation may be compared to the pre 1984 practice where K values are taken as 0.5 and 0.7 for tensile and compressive loading, respectively of an open ended pile. One ingredient that has lead to the current API recommendations was the conclusions from the API supported study of Dennis and Olsen (Ref. 14) which collected and examined the data for 66 closed andopen ended driven piles, tested in compression and in tension. The substantial increase of 60% in the K value for tensile loading has a major implication on the design of North Sea piles in sand. As a result, the Department of Energy and Fugro-McClelland supported a study in which the data base for pull-out tests on piles driven into sands was re-evaluated. This study which was reported by Lings (Ref. 15) consisted of re-interpreting the load test data and thesoil conditions for some 20 tests which were considered more reliable than others. The data base included proprietary data from a pull-out test on a 38m long conductor driven into dense to very dense North Sea sands. Lings I s conclusion was that the API (1984) method appears in general to give satisfactoryresults for piles up to 60 diameters penetration while at greater depths, the method seriously over predicts pile capacity. In a more recent work reported by Toolan and Ims (Ref. 16) the current API method for sand was also examined and shown