A review of modern approaches of oil spill modeling| A review of modern approaches of oil spill modeling |
 |
 |
|
KULYGIN V.V. Southern Science Center RAS, Russia Over the past three decades simplified empirical formulae contributed greatly in a rapid evaluation of the oil slick spreading and drifting.In the last years, simplified approaches of 1970s and 1980s were still popular in the oil spill modelling. The preference often was given to easier, but less accurate approach. Nowadays, due to recent developments in environmental and computational sciences, modern oil spill operational models can afford to use more accurate and physically relevant mathematical formulations. Oil spill modeling has many factors affecting the transport and fate of oil spills in the marine environment. These are the initial volume and the physico-chemical characteristics of the spilled oil, the characteristics of the spill region, the water composition and circulation under and around the oil spill, the wind field above the spill area, and several physicochemical and biological processes, such as: the spill spreading, turbulent diffusion, hydrodynamic and natural dispersion, evaporation, dissolution, emulsification, auto- and photo-oxidation, biodegradation, sinking/sedimentation, and resurfacing. All these processes and factors are interrelated and must be considered together to arrive at an accurate estimate of an oil's likely behaviour. Spills can originate from subsea pipelines (with or without presence of gas; from slow, long term) or virtually instantaneous releases of oil at the water's surface from tankers or platform operation. The characteristics of a spill source can greatly affect subsequent spill behaviour. So one developed models describes single oil spill source and another models can modeling different spill sources. Reed et al. (2006) estimate discharges from seafloor pipelines and not consider subsequent spill behaviour. SLROSM model can simulate all of enumeration above release scenarios. The initial slick area, thickness and properties are determined for various release scenarios and then the same basic oil fate models are used to predict the behaviour of the surface oil. SLROSM allows multiple slick releases. The behaviour of each slicklet is modeled separately because the slicklets may be subjected to different wind histories depending on their time of release. Many modern models rely on the work of Fay (1971) and Mackay (1980), but include some modifications. For example, Belore et al. include modification to account for oil viscosity changes and the development of a yield stress in the oil (pour point). As to vertical mixing Delvigne's (1988) oil entrainment model is often used.For more adequate description oil spill behaviour the three-dimensional models describing movement of oil slick on a surface and vertical interaction with a water column are considered. Earlier models (Fay, 1971; Mackay, 1980) predict temporal dynamics of the slick area, rather than slick shape, forcing researchers to postulate the shape to be circular, elliptical, or to fit to the "spillet" distribution pattern. Johansen (1982) and Elliott (1986) came out with a hypothesis that the observed slick elongation develops as the result of oil droplet sinking and resurfacing in the presence of shear within the upper layer of the water column. Recently, Tkalich and Chan (2002) considering dominant forces affecting the droplet formation and vertical distribution, developed a kinetic model of oil droplets vertical mixing due to breaking waves. Major characteristics of oil, waves and water column are combined into the "mixing factor", that is shown to be a generalization of the Delvigne's "droplet entrainment mass". Mixing factor includes oil viscosity, wave height dependence, the interfacial surface tension coefficient and oil density. MOSM (Tkalich, 2004) is one of the most full oil spill model. MOSM computes simultaneously six state variables (as in Fig. 1): thickness of slick on the water surface; concentration of dissolved, droplet and particulate oil phases in the water column; and concentration of dissolved and particulate oil phases in bottom sediments. MOSM represents a set of partial-differential equations. LNS equations governs the slick horizontal drift, associated with the combined action of wind and near surface current, and the slick spreading due to gravity-viscosity forces. The set of advection-diffusion equations describes transport of oil phases in the water column, identify oil transfer between the droplet, dissolved and particulate phases according to the prescribed kinetics. Following the approach, the mixing layer is identified beneath the slick to correct description of oil droplet vertical dynamic. While in this layer, the larger droplets may rise back to the slick, whereas smaller droplets may become permanently dispersed in lower layers of the water column. Being the most detailed, this model demands significant volume of the data, which not always are available. The choice of a degree of complexity depends from model's purpose, requirements and the available data. For example, depending on type of considered oil, evaporation is taken into account, or not; in mobile systems need a compromise between speed and resolution of vertical structure (Copeland G., Thian-Yew W., 2006). In conclusion notes, that modern knowledge of spilled oil behaviour is still very limited, the main purpose of the research is to develop a new up-to-date model skeleton and computational framework, able to include other relevant phenomena at the later stage. Reference OIL AND GAS OF ARCTIC SHELF 2006 (PROCEEDINGS OF INTERNATIONAL CONFERENCE)
Set as favorite
Bookmark
Email This
Hits: 19152 |
