| dc.description.abstract | Gas hydrate formation is a serious problem in the oil and gas industry, since its formation can plug wells and prevent production. The gas hydrate is a crystalline solid with a natural gas molecule surrounded by a cage of water molecules. It forms at high pressures and low temperatures. This is a problem for offshore gas wells, where the temperature is low in transport lines from well to the production facilities. Mono Ethylene Glycol (MEG) is commonly used as hydrate inhibitor. Classified as a thermodynamic inhibitor, this additive functions just as antifreeze in an automotive radiator.
When producing oil and gas there will in most cases also be produced some water, which can contain dissolved salts. These salts may precipitate and they tend to deposit on surfaces. Deposition of inorganic minerals from brine is called scale.
Generally MEG has the adverse effect of lowering the solubility of most salts. A common method to prevent corrosion in flowlines is to increase pH by adding basic agents (e.g. NaOH, NaHCO3) to the MEG stream. In such cases, carbonate salts are particularly troublesome since an increase in pH by one unit, will reduce the solubility by two orders of magnitude. Thus there will be a trade off between good corrosion protection (high pH) and scale control (low pH).
The aim of this work has been to develop a model that can predict mineral solubility in the presence of MEG. Experimental solubility data, together with thermodynamic data taken from literature, have been utilized to construct empirical functions for the influence of MEG on mineral scale formation. These functions enabled the expansion of an already existing aqueous scale model into a model valid for water+MEG mixed solutions. The aqueous scale model combines an equation of state (gas+oil phase) with the Pitzer ion interaction model (water phase) to describe the multiphase behaviour of gas-oil-water systems. This work describes how MEG has been introduced into the water phase model.
The general idea is that the activity of a specie, ί, is given as its concentration, m, times the activity coefficient, γ , which is divided in two parts. γS shall take care of the “Salt effects”, and γN the “MEG effects”;
ai=miγi=miγs iγNi
γS is calculated by the Pitzer model, as if the solvent was water, and consequently has the same numerical value regardless of the MEG concentration in the water+MEG solvent. γN is empirically fitted from solubility data and is obviously a function of MEG concentration. γN may also be dependent on temperature and ionic strength.
The pressure dependence of γN has not been investigated in this work. All equilibrium constants, K°, are independent of MEG concentration. Theoretically this corresponds to a pure water standard state. This modelling approach has the advantage that it gives a simple and robust model with reasonable extrapolations outside the range of experimental data. Practically it turns out that the effects of temperature, and that of dissolved species (ionic strength) are almost the same in water+MEG solutions as in water. In such a case a good approximation will be to let the γN term merely be dependent on MEG concentration. It has been shown that γN for some systems is a function of both temperature and ionic strength in addition to MEG concentration.
The mathematical functions used for curve fitting γN were generally arbitrarily chosen polynomials, meaning that they do not have any physical/theoretical basis.
The resulting model has good flexibility and can do exactly the same type of calculations as the aqueous model. It can handle MEG concentrations of up to 99 weight % in the solvent. MEG concentration is commonly specified in the water phase, but the model also accepts MEG input in the gas or oil phase. For conditions encountered in oil and gas transport pipelines and at well heads, the model should function well. It is empirically fitted from solubility data, generally covering the range 0-100°C and 0-100% MEG in the solvent. Hence if the model predicts precipitation of a salt, the ionic strength is normally comparable with the value in the data used for fitting the model. In e.g. a MEG regeneration boiler, however, the temperature is high, and/or several highly soluble species like Na+, K+, CO32-, Cl- are present yielding a very high salinity. The model is therefore not suited for calculations at such conditions.
MEG influences the phase distribution of gases. This effect has been included for CO2, H2S, CH4 and the most common organic acids found in the oilfield. CO2, H2S, and the organic acids also have MEG dependent dissociation equilibria.
The scale forming minerals included in the model are:
→ CaSO4, CaSO4·2H2O, BaSO4 and SrSO4
→ CaCO3, FeCO3, BaCO3, SrCO3, and 3MgCO3·Mg(OH)2 ·3H2O
→ NaCl and KCl
→ NaHCO3 and KHCO3,
→ NaAc and NaAc·3H2O
→ Na2CO3, Na2CO3·H2O and Na2CO3·10H2O
→ K2CO3 and K2CO3·1.5H2O
→ FeS
→ Mg(OH)2,
Much new data have been gathered for the water+MEG system, mainly concerning the first dissociation constant of CO2, the solubilities of the carbonates; CaCO3, BaCO3, SrCO3 and 3MgCO3·Mg(OH)2·3H2O as well as the sulphates, CaSO4 and CaSO4·2H2O. These experiments were confined to 20-80°C and ionic strengths of 0-0.7mol/kg.
Hydromagnesite (3MgCO3·Mg(OH)2·3H2O) has been included as the only magnesium carbonate mineral in the scale model. Hydromagnesite is actually a meta-stable phase, but the thermodynamically stable magnesite (MgCO3) has been omitted due to its formation being kinetically inhibited. Magnesite of sufficient purity for solubility investigations was not available from commercial suppliers. It was therefore necessary to synthesize it in the laboratory. A new method for synthesizing magnesite from hydromagnesite at atmospheric conditions has been suggested. Mono Ethylene Glycol
(MEG) is used to lower the solvent vapour pressure at temperatures above 100°C.
The MEG concentration of an unknown sample is often measured using Gas chromatography (GC). This is an accurate method but has the disadvantage that the sample very often has to be shipped to an external laboratory. A new method for prediction of MEG concentration that is fast, easy and inexpensive has been developed. Values of density, conductivity and alkalinity of an aqueous solution, are used to estimate both MEG and salt contents. The method is valid in the whole concentration interval of 0 to 100 wt% MEG and with ionic strengths from zero to the solubility limits of NaCl and NaHCO3. At intermediate MEG concentrations (40 to 90wt %) the accuracy is regarded as ±2 wt % for MEG content determination. The main limitation is that NaCl and/or NaHCO3 must be the dominating dissolved salts.
pH is an important parameter in carbonate scale prediction. This work summarizes the theoretical foundation and proposes how to work with pH in water+MEG solutions. A pH electrode calibrated only in common aqueous standard solutions, gives a measured value denoted pHmeas in this work. pHmeas is not reproducible in water+MEG solutions. Calibration also in 0.05m KHPh (Potassium Hydrogen Phtalate) solutions with certain MEG concentrations, gives the calibration value; ΔpHMEG. The actual pH, which is reproducible, can thereafter be found from pHmeas as;
pH = pHmeas + ΔpH MEG+ ΔpHSalt
ΔpHMEG has to be determined once for each electrode. ΔpHSalt adjusts for the salt/ionic strength impact on the electrode and is only important at ionic strengths above ~0.5mol/kg. | nb_NO |
