Modelling the dynamic embedment of seabed pipelines

Abstract

The as-laid embedment of a seabed pipeline is an important design parameter. As a pipe is laid on the seabed it oscillates, owing to vessel motion and hydrodynamic loading of the hanging pipe. This movement significantly increases the pipe embedment beyond the theoretical value related to the static pipe weight, even when corrected for any stress concentration caused by the hanging catenary. Dynamic lay effects are either ignored in practice, or are accounted for by scaling up the static embedment by an empirical factor, leading to significant uncertainty in this important design parameter. A series of centrifuge model tests has been conducted using two clays – kaolin and a high-plasticity natural clay – to simulate the dynamic embedment process. The results indicate that only a few cycles of small-amplitude oscillation (±0·05D) are required to double or triple the pipe embedment, owing to the combined effect of lateral ploughing and soil softening. In these experiments the pipe embedment increased to up to eight times the static embedment after 100 cycles of motion, which represents a typical lay process. A model is proposed for the cycle-by-cycle embedment of a pipeline under a given sequence of small-amplitude oscillations at a given applied vertical force. The trajectory of the pipe movement is assessed using a flow rule derived from plasticity-based yield envelopes. The effect of soil remoulding is explicitly captured by linking the accumulated disturbance to the decay in soil strength. Using input parameters derived from theoretical considerations and T-bar penetrometer tests, the model captures the essential features of the dynamic embedment process. With modest optimisation of the model parameters, the mean discrepancy between the calculated and measured embedment is only 12% for both clays. The ultimate states predicted by this cycle-by-cycle model also provide a rough estimate of the maximum pipe embedment for fully remoulded conditions, which include some degree of water entrainment caused by the lay process, evident in the optimised parameters. This ultimate embedment is governed by the remoulded soil strength and the pipe weight (augmented by any stress concentration). The amplitude of the cyclic motion affects the rate of softening, and hence the rate of settlement. This model provides a framework for assessing the as-laid embedment of seabed pipelines on a more rigorous basis than current practice.