The gas hydrates are solid compounds of clathrate type which can be formed starting from cold water and hydrocarbon gas molecules under pressure. These conditions are met in certain oil conduits and can lead to a problem of production. Indeed, the oil effluent which leaves a well of production always contains light water and hydrocarbon molecules (methane, ethane, propane) suitable form a gas hydrate. The methane hydrates are not naturally present in the layers of production because the temperature is too high (until 200°C). On the other hand, the oil fluid cools at the time of its transport in a control, either because control is localized in a particularly cold zone, or because control is underwater, by the contact with cold water. It can then create hydrates being likely to block the conduits. To prevent their crystallization, the current tendency is to couple three types of approaches: insulation of the conduits, injection of additive at the time of the critical phases, reheating of control by hot water circulation at the time of accidental stopping. This thesis takes part in the modeling of the flows after formation of hydrates. It is not thus a question of preventing crystallization but of being interested in rheology of the flow after formation of crystals. The long-term objective is to identify the origin of the transportability of the purées of hydrates under the influence of additives known as "anti-binders". The mechanisms of crystallization indeed are very often coupled: germination, growth, agglomeration, attrition... The comprehension of the mechanism of action of an additive is thus a complex spot, more especially as crystallization is closely related to the physical system in which it develops. The studies seeking to identify the mechanisms of crystallization, for then including/understanding the effects of additives all (or practically) were carried out in closed engines and simple systems (eau/gas) (Herri (1996), Pic (2000)...). Conversely, tests of validation of additives were carried out on loops control representing a real flow, therefore complex. Our work is thus halfway of the two preceding approaches. It is a question of approaching the geometrical conditions of an oil flow (pilot buckles) while preserving a simple system (eau/dodécane) with for objective on the long term identifying the coupling: geometry/crystallization/influence of the additives. The experimental device (height 12m, width 3m, length 6m) carried out within the framework of this thesis is a pilot loop of circulation reproducing certain conditions of the flow of an oil fluid (emulsion water in oil) in a subsea pipe, i.e. under strong pressure [ 1-10 MPa ] and low temperature [ 0-10°C ]. The various parts of this instrument are:
The serpentine being rolled up on 3 levels
The tube going up is a riser. With the base of this tube is injected methane in order to reduce the column of fluid to create an effect elevator: gaslift.
The separator located at the top of the riser separate by gravity the gas part, of the liquid part (water, oil) which goes down again in a tube parallel with the riser towards the loop of circulation
The system of recompression of gases recovers gases of the separator to reinject them with the bottom of the riser after increase in the pressure. This experimental system is composed in fine of two loops circulating on themselves: a loop liquidates and a loop gas. These two loops share a common section made up of the riser and separator. This device made it possible to make rheological studies on the phase only continues (dodecane) according to the pressure of methane and on emulsions containing various water contents and of additives. Studies concerning the crystallization of the methane hydrates within the emulsions were carried out by considering the influence of the water content then that of the content of additive on the apparent viscosity of dispersions thus formed. We finally propose a modeling connecting crystallization to the rheological behavior.