Optimizing drivability predictions for piles in very dense sand

Due to growing demand of offshore wind energy in Europe, large diameter monopiles are being installed throughout the North Sea. GeoSea, Deme’s specialist in offshore solutions, during installation of monopiles in the North Sea, encountered lower soil resistances in very dense sand layers than what was expected from results of offshore geotechnical investigations. This paper aims to study plausible physical phenomenon happening in those very dense sand layers and to propose a way to predict the behavior of those layers for future drivability analyses. In order to do so, first a review of the literature examines the basic principles of pile drivability and drivability analyses. To conduct drivability analyses, the commonly used program GRLWEAP is adopted. This computer program is based on the wave equation’s model proposed by Smith more than 50 years ago. This paper proposes to examine the effects of different parameters involved in a drivability study. Due to the considerable number of piles to be analyzed, processing of the soil investigation data is done via automation. This is accomplished via an Excel spreadsheet using commonly used CPT correlations and Robertson’s work on soil type behavior. In this spreadsheet, it is implemented the Alm & Hamre (2001) method that calculates one of the most important parameters implicated in a drivability study, the static resistance to driving (SRD). Drivability analyses are first conducted using the author’s recommendations. However, the procedure being time-consuming and the number of piles to be analyzed being numerous, an approach to fasten the procedure is examined. This approach allows to conduct drivability analysis in one step using a reviewed and updated method of Schneider and Harmon. Then, a first drivability study is conducted on three locations where instrumentation at the pile head measured blow counts, compressive stresses and transferred energy to the pile. Wave equation results confirm the effect noticed in the sand layers by GeoSea. Back analyses were then conducted using an iterative process to fit the back calculated profiles to the measured ones. Several ways of handling the problem are tested during the iterative process but only one method fitting the measured data is mentioned in the thesis. To get best fit results, it is proposed to limit cone resistance values obtained by soil investigations. The second part of the thesis tries to bring clarity on the physical phenomenon happening in the sand layers. Two ideas (cementation and slab effect), and their potential implication on soil quake values, are analyzed, but results are inconclusive, and no conclusion can be made about the behavior of the sand layers. Ultimately, the proposition of considering limiting qt values is tested on multiple locations where pile installation presented the problem noticed by GeoSea. According to CPT profiles,16 piles in total presented penetration in very dense sand layers, so that the proposed procedure is applied for all these piles to test it. Results of the procedure show good fit with the measured blow count profiles for the very dense sand layers. But poor fit is obtained on the surrounding clay layers. Finally, in conclusion, we propose recommendations regarding the developed procedure, discuss the different results obtained along the thesis and open the path for future work.