Abstract

Through complementary experimental and numerical efforts, the present paper aims to make a significant contribution to the overall understanding of backfilling processes beneath submarine pipelines. For this purpose, we aim to simplify the experimental backfilling process to an elementary two-stage process: (1) initial scour induced by a pure current, followed by: (2) backfilling induced by pure waves. A steady current is introduced via a re-circulating pump, and is kept constant with a cross-sectional velocity of V = 0.48 m/s until an initial equilibrium scour depth, S0, is reached. Then, the current is stopped and waves (characterized by their Keulegan-Carpener number KC and Shields parameter θ)are introduced to initiate the backfilling process, which is maintained until a new equilibrium scour depth, Sf, is reached. The time at which waves are introduced will be denoted as t = 0. For the backfilling process both regular and irregular waves are used during the experiments. As a demonstration of the initiated two-stage (scour followed by backfilling) process, bed profiles based on video recordings from a case having KC = 9.7 and ϴ = 0.195,are depicted at selected stages in Figure 1. Figure 1(upper left) depicts the current-induced equilibrium scour hole in the near vicinity of the pipe at t = 0, with the profile approximated as the dashed red line. Similarly, Figure 1(upper right) depicts the new equilibrium scour profile (approximated as the full blue line) that has developed under wave-induced backfilling, corresponding to t = 60min. To ease comparison, the dashed red and full bluelines from these plots are additionally combined onto Figure 1 (bottom).The experimental campaign has additionally been complemented with similar numerical simulations (using regular waves), based on a fully-coupled hydrodynamicand morphodynamic CFD model (Jacobsen et al., 2014),extending previous pipeline scour-related applications ofFuhrman et al. (2014) and Larsen et al. (2016). Comparison of the numerical and experimental results demonstrate the ability of the CFD model to reasonably simulate the current-to-wave backfilling process, both interms of the achieved new wave induced equilibriumscour depths as well as the corresponding backfilling timescales. Figure 2 depicts a summary of both experimental and numerical backfilling time scale Tb versus Shields parameter θ. As can be seen, both experimental as well as numerical results match the regression equation:T"sub"b"/sub"=0.3 θ-"sup"5/3 "/sup"quite closely (solid line in Figure 2).