Condensed from paper No. 19, presented at NACE Corrosion/96
Reproduced with permission from NACE International.

Experiments have been performed in a 10 cm ID horizontal pipeline to observe the effect of temperature, carbon dioxide partial pressure, flow, and oil/water composition, on corrosion in horizontal multiphase slug flow, using a low viscosity (2 cp) oil. Corrosion rates have been measured for water cuts ranging form 20 to 100%, carbon dioxide partial pressures of 0.27, 0.45 and 0.79 MPa, temperatures of 40, 60, 70, 80 and 90 C and Froude numbers of 6, 9 and 12. The average void fraction as well as the pressure drip across the slug is found to increase with an increase in Froude number and carbon dioxide partial pressure. A model to predict corrosion rates for low viscosity oils has been established which relates the corrosion rate to the pressure gradient, temperature, carbon dioxide partial pressure and water cut.

When the oil fields are located in remote areas such as Alaska or subsea, it is not practical to separate the oil/water/gas at the well site. Hence, it is a common practice to transport the oil/water/gas mixture from several wells to a central gathering station through large diameter pipelines, where separation takes place.

As the well ages, enhanced methods of oil recovery involving the injection of carbon dioxide and water are used. This helps to maintain the pressure within the reservoir. However, some of the carbon dioxide and brine flows with the oil and gas. This multiphase flow of oil-water-gas mixture creates a number of corrosion problems in the pipelines. Many oil wells operate at high water cuts, as high as 80%. Also, in the presence of water, the added carbon dioxide forms weak carbonic acid. This acid being corrosive in nature, causes higher corrosion rates in carbon steel pipelines. The oil and gas mixture may also contain waxes, hydrates, hydrogen sulphide and sand.

Efird, Wright, Boros and Hailey (1993)6 performed experiments with three different systems, 2.54 cm and 9 cm pipe diameter loops, jet impingement, and rotating cylinder electrode method. They established a correlation between corrosion rate and wall shear stress. This correlation is given by the following equation:

where, R_COR = corrosion rate in mm/year

t _{w =} wall shear stress in N/m²

a & b = constants

This equation is valid only for brine and different values of a and b are needed if it is to be applied to other systems. The value of the coefficient "a" varied with the temperature, carbon dioxide partial pressure and the type of flow. These tests also provided a comparison between pipe flow, rotating cylinders and the jet impingement technique. It was found that the pipe flow correlated better with the jet impingement technique, while the results from rotating cylinders grossly underestimated the corrosion rates

Kanwar and Jepson (1994) performed corrosion studies in a 10 cm diameter flow loop under full pipe flow conditions, at carbon dioxide partial pressures up to 0.79 MPa, temperature of 40 C with two oils of viscosities 2 and 18 cp and ASTM water. Based on the concept introduced by Efird et al.(1993), they proposed the following predictive model for corrosion rate:

where,

R_COR = corrosion rate in mm/year

P = carbon dioxide partial pressure in MPa

_w = wall shear stress in N/m²

b & c = constant exponents with values of 0.1 and 0.83 respectively

k = constant (mm/year)(MPa)^-0.83 (N/m²)^-0.1 (K)^-1

This relation is valid for full pipe flow of low viscosity oils with concentrations up to 60% oil and a temperature of 40 C. Later, Kanwar (1994)8 carried out similar experiments at temperatures of 30, 33, 50 and 60 C to determine the effect of temperature on the corrosion rate. He found that the coefficient "k" in above equation 2, could be represented as a function of temperature in the following manner:

where,

k₁ = constant with value 416649 (mm/year)(MPa)^-0.83 (N/m²)^-0.1(K)^-1

T = temperature in Kelvin

However, this equation cannot be applied at temperatures higher than 60 C, since it does not account for the formation of protective iron carbonate scales at these higher temperatures.

The present work focuses on the effect of flow, temperature, carbon dioxide partial pressure and oil composition on the corrosion rate in horizontal multiphase slug flow.

The fluids used were a refined oil with a viscosity of 2 cp at 40 C and density of 800 kg/m3 and ASTM standard sea water. Carbon dioxide is used as the gas phase.

The experiments were performed at carbon dioxide partial pressures of 0.27, 0.45 and 0.79 MPa, temperatures of 40, 60, 70, 80 and 90 C and a Froude number of 6, 9 and 12 (corresponding to slug velocities of 3, 4.5 and 6 m/s respectively). Water cuts of 100, 80, 60, 40 and 20% were used.

It is seen that at each of the carbon dioxide partial pressures and Froude numbers, the corrosion rate increases with an increase in temperature. No maximum in the corrosion rate is seen at any of the temperatures studied.

It is seen that the oil fraction is almost uniform across the pipe diameter indicating that the liquid phase is well mixed under slug flow conditions. The oil is now the continuous phase at this percentage of oil in the liquid. Hence, with increase in oil composition, less water contacts the pipe surface leading to decrease in corrosion rate. Also it can be seen that the void fraction here at the bottom of the pipe decreases with increased in oil composition. This results in a decrease in the intensity of turbulence, resulting in the decrease in the corrosion rates.

Since, the pressure drop across the slug is a measure of the shear and turbulence at the bottom of the pipe, the corrosion rates at each temperature were plotted against the experimentally obtained pressure gradients. It is found that, under almost all conditions the value of the exponent remains constant and its value is found to be 0.30 0.05. Similar results were obtained at lower water cuts of 80, 60 and 40%. So the ollowing relation between the corrosion rate and the pressure gradients is proposed:

CR = c ^0.3 v^0.6

The equation for corrosion rate as a function of carbon dioxide partial pressure, water cut and pressure gradient becomes:

CR = K₁ T exp

The value of the coefficient k(T) from equation is plotted against temperature. A regressional analysis of the values of k(T), was carried out to evaluate the activation energy constant 'E'.. The resulting equation defines the value of the constant k1 and the activation energy. These values are,

Thus the final expression for the corrosion rate as a function of pressure gradient, water cut, carbon dioxide partial pressure and temperature is given below.

CR = 31.15 v^0.6 P_CO2^0.8 T e

For the low viscosity oil tested, the corrosion rate increased with an increase in temperature over the entire range studied, at every Froude number, carbon dioxide partial pressure and water cut. No maximum in the corrosion rate is observed at any of the temperatures studied. The high levels of shear and turbulence at the bottom of the pipe, encountered in slug flow, strip away the protective film of corrosion products formed on the pipe wall resulting in high rates of corrosion.

At each temperature, carbon dioxide partial pressure and water cut, the corrosion rate increased with an increase in Froude number. Also, with increase in Froude number, there was a increase in the void fraction at the bottom of the pipe as well as the average void fraction across the pipe diameter. This was manifested in an increase in the pressure drop across the slug indicating that the intensity of shear and turbulence increased with increase in Froude number.

The corrosion rate decreased with an increase in oil composition from 0 to 60%. The corrosion rate reduced to negligible values for an oil composition of 80%. This was due to the transition from a water continuous phase to an oil continuous phase. No maximum in the corrosion rate was seen at any of the oil compositions studied.

The oil fraction across the pipe diameter is more or less uniform across the pipe diameter indicating that the liquid phase is well mixed under slug flow conditions.

For slug flow, the pressure gradient is a measure of the shear and turbulence at the bottom of the pipe. A model to predict corrosion rates has been established which shows that the corrosion rates are dependent on the temperature, carbon dioxide partial pressure, pressure gradient across the slug and water cut. This model can predict corrosion rates up to water cuts of 60% for a low viscosity oil, carbon dioxide partial pressures up to 0.79 MPa and temperatures up to 90 C.