Corrosion

The prediction of CO₂ corrosion in oil and gas production systems involves assessing corrosion rates in the absence of corrosion control measures, using one of three established CO₂ prediction models:

• deWaard 1991 Model
• deWaard 1995 Model
• Norsok M506 Model

The corrosion rates calculated from the corrosion models represent uninhibited corrosion. The corrosion allowance will be determiend by combining uninhibited and inhibited corrosion.

Design Parameters

The key design inputs for corrosion prediction include temperature, total system pressure, CO₂ partial pressure, pH value, flow velocity, and shear stress. Some additonal inputs are required for the deWaard models. These parameters include the condensation factor, scale factor, liquid velocity, and hydraulic diameter. Wall shear stress is an additional input for Norsok M50 model. It can be provided directly or calculated using flow properties. To determine the corrosion allowance, which combines uninhibited and inhibited corrosion, the following parameters are required: design life, inhibited corrosion rate, and inhibitor availability.

deWaard 1991 Model

deWaard 1991 model is a predictive model for CO₂ corrosion in wet natural gas pipelines. Initially, the deWaard-Milliams equation and the corresponding nomogram corroion rate has one temperature-dependent term and one CO₂ partial pressure-dependent term, as shwon in Equation \eqref{eq1}. \[ logV_{nomo} = 5.8 - {1710 \over T} + 0.67log \left(pCO_2\right) \tag{1}\label{eq1} \]

where

• \(V_{nomo}\) = nomogram corrosion rate (mm/year),
• \(T\) = temperature (ºK), and
• \(pCO_2\) = partial pressure of \(CO_2\) (bar).

The deWaard 1991 model proposes modifying this conservative prediction by multiplying V_nomo by several correction factors:

• Effect of total pressure \(F_{system}\)
• Effect of high temperature protective films \(F_{scale}\)
• Effect of Fe++ and and pH \(F_{pH}\)
• Condensation factor \(F_{cond}\)
• Effect of glycol \(F_{glyc}\)

An increase in CO₂ partial pressure will lead to a higher corrosion rate. However, as pressure increases, the nonlinearity of natural gas behavior becomes more significant. Therefore, CO₂ fugacity (fCO₂), rather than CO₂ partial pressure, should be used corrosion rate calculation. To account for this effect, a pressure factor F_system is proposed and appplied.

As temperature increases, a protective film begins to form, leading to lower corrosion rates. At a certain high temperature, the corrosion rate reaches a maximum. This temperature is referred to as the scaling temperature, which is likely dependent on flow rate. The effect of high-temperature protective films is represented using a scale factor. The temperature at which F_scale = 1 or log(F_scale) = 0 is referred to as the scaling temperature, at which the corrosion rate reaches its maximum.

Depending on temperature and CO₂ pressure, the corrosion of steel in the CO₂/water system can lead to an increase in Fe++ concentration. The contamination of the CO₂ solution with corrosion products reduces the corrosoin rate. Without the presence of corrosion products, much higher corrosion rates are possible. To account for this effect, a correction factor for pH shift caused by the presence of dissolved Fe++ is introduced.

Corrosion rates of steel exposed to a condensing water phase in a CO₂ atmosphere decrease rapidly over time. On average, these rates can be reduced to 1/3 or even 1/10 of the nomogram corrosion rate, depending on the rate of water condensation. For wet gas transport, when condensation rates are typically below 0.25 g/[m²s], it is conservative to use F_cond = 0.1.

Glycol is often added to wet gas pipelines to prevent the formation of hydrates. The presence of glycol also reduces the corrosion rate. This effect of glycol is represented by a factor of F_glyc.

Due to varying design concepts regarding corrosion models, inhibitor availability philosophies, corrosion control methodologies, and national standards, significant differences in corrosion rates and corrosion allowances are expected. The selected prediction model(s), design parameters, and corrosion rates should be agreed upon and verified with the end user to incorporate field experience.

deWaard 1995 Model

The deWaard 1995 Model is based on extensive data from numerous experiments conducted in a high-pressure test loop under strictly controlled environments and flow conditions. The CO₂ corrosion rates from these experiments were fitted to the de Waard 1995 Model as a semi-empirical equation. This model combines the flow-independent kinetics of the corrosion reaction with the flow-dependent mass transfer of dissolved CO₂ using a resistance model. \[ {1 \over V_{cor} } = {1 \over V_r} + {1 \over V_m} \tag{2}\label{eq2} \]

where

• \(V_{cor}\) = corrosion rate (mm/year),
• \(V_r\) = reaction rate (mm/year), and
• \(V_m\) = mass transfer rate (mm/year)

In the deWaard 1995 Model, the effect of pH is incorporated into the reaction rate term V_r. By including the mass transfer term V_m included, the deWaard 1995 Model achieves better correlation with test data, accounting for the influence of liquid flow velocity on CO₂ corrosion.

Norsok M506

NORSOK M-506 presents a method for calculating corrosion rates in hydrocarbon production and process systems where the corrosive agent is CO₂. The NORSOK M-506 model is an empirical corrosion rate model for carbon steel in water containing CO₂ under varying temperatures, pH levels, CO₂ fugacities, and wall shear stresses. It is based on flow-loop experiments conducted at temperatures ranging from 5 ºC to 160 ºC.

The pH value and wall shear stress are the primary parameters for the Norsok M506 Model. The calculation module offsers options to input these parameters directly or to calcuate them using other input parameters. For accurate wall shear stress calculations, additional parameters such as mixture density, velocity, and viscosity are required. Users are advised to use the calculated shear stress cautiously, especially in the presence of obstacles and other geometric changes in the flow that are not accounted for in the calculation.

It is important to note that pH should be measured under in situ conditions rather than after depressurization and atmospheric exposure. The pH value can also be calculated for both condensed water and formation water (produced water) based on chemical reactions and equilibrium constants. These equilibrium constants can be determined using parameters such as temperature, pressure, total acetic acid, alkalinity, and ionic strength.

Inhibitor Availability

The uninhibited corrosion rates are determined using the above corrosion models. The inhibited corrosion rates, which should reflect actual field conditions, will be provided. Inhibitor availability depends on the planned corrosion management program, including corrosion monitoring and inhibition strategies. An inhibitor availability of % is used. The corrosion allowance, taking into account both inhibited and uninhibited periods over the lifetime, is determined by the equation shown below. \[ CA = \sum_{i=1}^n \left[ {A \over 100} \cdot CR_{in} + \left( { {100-A} \over 100} \right) \cdot CR_u \right]_i \tag{3}\label{eq3} \]

where

• \(CA\) = corrosion allowance (mm),
• \(A\) = availability of corrosion control [%],
• \(n\) = number periods with constant conditions,
• \(CR_{in}\) = corrosion rate with corrosion control (mm/year), and
• \(CR_{u}\) = corrosion rate witout corrosion control (mm/year).

Parameter Study

Parameter studies are also conducted to evaluate the influence of various factors on corrosion rates. One example of the calculated corrosion rates with varying CO₂ mole% are shown in Figure 1 for different corrosion models. The figure displays three lines for a constant temperature of 100.0 ºC. The base cases, marked as spots, correspond to a CO₂ mole% of 2.0% CO₂ (mole%) at this temperature.

Another example is the calculated corrosion rates with varying temperatures (scale effect) as shown in Figure 2 for different corrosion models. The figure plots three lines for a constant CO₂ partial pressure of 0.2 bar. The base cases, indicated by spots, correspond to a CO₂ partial pressure of 0.2 bar and a temperature of 100.0 ºC.

References

1. C. de Waard, U. Lotz, and D.E. Milliams, "Predictive Model for CO₂ Corrosion Engineering in Wet Natural Gas Pipelines", CORROSION (1991), Paper 577.
2. C. de Waard, U. Lotz, and D.E. Milliams, "Influence of liquid flow velocity on CO₂ corrosion: A semi-empirical model", CORROSION (1995), Paper 128.
3. Norsok Standard M-506: 2017 - CO₂ Corrosion Rate Calculation Model.
4. Norsok Standard M-001: 2014 - Material Selection.