On the modelling of installation effects on laterally cyclic loaded monopiles

Abstract

Monopiles are being used more extensively for offshore wind turbine foundation. Current guidelines on the construction of fixed offshore foundations still base the design of laterally loaded piles in sand on empirical data (p-y curves), originating from the oil and gas industry, from field tests executed in 1974 on 0.61 m diameter piles. However, large diameter monopiles can be considered short and rigid, which rotate rather than bend when subjected to lateral loads. As a result, numerous experimental studies have been performed on the response of monopile foundations by means of a geotechnical centrifuge. During these tests there was no consistency in installation conditions (ether at low stress levels (1g), elevated stress (Ng) levels or pre-installed), partly because the effect of the pile installation is still not completely understood. Research into the behaviour of these large open-ended piles generally does not examine the installation effect on lateral capacity. This research contributes to (i) investigate the effect of pile installation on the lateral response of an open-ended pile in sand and (ii) increasing reliability in the interpretation of existing and future centrifuge research on the investigation of lateral cyclic large strain deformations. The main objective of this thesis is to investigate the effect of monotonically jacked open-ended pile installation, at low and elevated stress conditions, on the large strain lateral soil-pile response during two-way cyclic loading. A series of tests was executed to compare the differences on the lateral pile response from monotonic jacked pile installation at low stress levels (1g) and at elevated stress levels in the geotechnical centrifuge (Ng). In order to study the effect of monotonic open-ended pile installation on the cyclic lateral capacity a novel actuator, which allows installing the pile in-flight and subsequently loading the pile laterally without interrupting the test, was developed. The load mechanism on the free pile head was designed in a way that no bending moment was transferred to the pile head by means of a hinged connection between pile and actuator. The load mechanism - instrumented with strain gauges - was calibrated for stiffness by means of static loads in the laboratory and for hysteresis at 1g and Ng in the geotechnical centrifuge. The brass model pile was designed to properly scale the lateral bending stiffness and prevent plugging during installation. In order to investigate influence of the soil state on installation effects a total of two sets of tests have been performed with varying installation conditions and relative densities. The first set (76g and ID = 60 +/- 3 %) was executed in duplex to determine the consistency and the accuracy of the preparation method. In the second set (48g) the relative density was varied (60 +/- 3 % and 80 +/- 3 %). The second set of tests provided a clear insight into relation between load and displacement during two-way cyclic loading (load - displacement loops). The results indicate that elevated stress installation of an open-ended pile has a small positive (1 - 5%) effect on the lateral capacity. This effect is mainly visible during primary lateral loading of the soil and decays with the number of load cycles. Hence for open-ended model piles the installation effects are negligible. Moreover, these tests indicate that in all cases the stiffness and lateral capacity increase with the number of load cycles, stress level and/or initial density. These gains in capacity and initial stiffness are much more substantial than the differences found between 1g and Ng pile installation. The results in this study present an incremental advance in fully modelling the installation effect and subsequent lateral cyclic loading of monopiles.