Numerical Modelling and Experiments on Sea Spray Icing

Abstract

Oil and gas activities in the cold seas are endangered by icing. Icing may occur when water spray or moisture is deposited on a vessel or offshore structure above sea level and the air temperatures is below the freezing temperature of water. Ice accumulation can block rescue equipment and doors, and clog ventilation systems, which may increase the risk of explosion as result of volatile gas accumulation. Icing can be caused by supercooled fog, freezing rain, falling snow and freezing sea spray. The freezing sea spray caused 80 – 90% of all offshore icing incidents and is the focus of this study. To take precautions against icing, it is important to understand the physics of icing and to model it. The rate of ice accretion is mainly defined by the spray flux and heat transfer, and both must be accurately predicted. Existing icing models, e.g., ICEMOD and RIGICE04, simplify the structure, subdividing it into cylindrical and flat components. In these models, airflow around a component is assumed to be unaffected by other parts of the structure, and the heat transfer is approximated using empirical relations. In reality, however, the airflow field is complex. Upwind components can create wind shadow regions or regions of accelerated flow in front of downwind components; this changes both the spray flux and the heat transfer, thus making ICEMOD and RIGICE04 inadequate. A research subject of this PhD study was the heat transfer and spray flow around a structure in a real airflow. The full-scale measurements in the literature are limited and it is expensive to perform them. Therefore, for a preliminary answer the question was addressed using computational fluid dynamics (CFD), which is capable of predicting the spray flow and heat transfer around a structure with any shape; however, the accuracy of CFD is uncertain. Existing models of sea spray icing, e.g., ICEMOD and RIGICE04, neglect heat flux into the accreted ice and assume that air cooling is directly spent to freeze the water film on the ice surface. This assumption is good for steady ice growth. However, it is also used in modelling icing caused by periodic sea spray. This study proves numerically and experimentally that the heat flux into the accreted ice generated by freezing must not be neglected. The main contributions of this work are as follows: The study develops the marine icing model, MARICE, which uses CFD to calculate spray flux and heat transfer, and models water film motion on any arbitrary surface. · The study shows that Reynolds-averaged Navier-Stokes (RANS) models should be used with care in icing simulations: the collision efficiency is well predicted upstream flow separation point by any RANS turbulence model; however, the wake and separation of the flow is modelled poorly. · The study develops and validates a new model of ice growth caused by periodic sea spray, which accounts for the heat conduction inside the accreted ice.