Corrosivity Assessment Model

CO₂/H₂S corrosion in oil and gas production environments represents one of the most important areas of corrosion research. It is so because of the criticality of the need to assess corrosive severity as a means to ensure safe utilization of steels, which have wide application in just about every sphere of oil and gas production and refining. Even though CO₂/H₂S corrosion and concomitant mechanisms have been areas of significant work over the last thirty years, there still exists a need to accurately predict corrosivity of CO₂/H₂S environments from a stand point of defining limits of use for carbon steels. Even though numerous predictive models have been developed and are being developed, most of the available predictive models tend to be either very conservative in their interpretation of results or focus on a narrow range of parametric effects, thereby limiting the scope of the model's application in realistic assessment of corrosivity and corrosion rates. Often times, data required by the models are often not easily accessible or available to the operators who need to employ the model, thereby limiting the applicability of the models to situations of reduced practical importance. In this context, the issue of corrosivity assessment for carbon steels can be re-stated in terms of the following critical requirements:

• Develop a predictive model that utilizes commonly available operational parameters
• Utilize existing lab/field data and theoretical models to obtain realistic assessments of corrosivity and corrosion rates
• Develop a computational approach that integrates both numerical (read "lab trends") and heuristic (field data and experience) information and knowledge about corrosivity prediction.

The primary variables in in the Predict model are the acid gases CO₂ and H₂S that contribute to the typically acidic pH found in production environments. The model uses the widely accepted de Waard - Milliams[2] relationship for CO₂ corrosion for an initial determination of CO₂-based corrosion rates. However, the effective CO₂ partial pressure in the system is not based on the operating partial pressure but one obtained from the system pH. This rate is further refined to account for the presence of H₂S, corrosion products, temperature effects etc. The underlying idea here has been to develop a prediction model that accurately represents the state-of-the-art in theoretical analyses as well as parametric correlations based on lab and field data.

While there have been several studies focusing on the exact mechanism of metal dissolution in CO₂ containing waters, the efforts of De Waard and Milliams and others[2,3,9] present a commonly accepted representation wherein anodic dissolution of iron is a pH dependent mechanism as given by Bockris, the cathodic process is driven by the direct reduction of undissociated carbonic acid. These reactions can be represented as,

Fe ----------> Fe⁺⁺ + 2e^- (Anodic reaction)

H₂CO₃ + e^-----> HCO₃^- + H (Cathodic reaction)

The overall corrosion reaction can be represented as,

Fe + 2H₂CO₃ ---> Fe++ + 2 HCO₃^- + H₂

The build up of the bicarbonate ion can lead to an increase in the pH of the solution till conditions promoting precipitation of iron carbonate are reached, leading to reaction given below:

Fe + 2HCO₃^- ---> FeCO₃+ H₂O+CO₂

Iron carbonate solubility, which decreases with increasing temperature, and the consequent precipitation of iron carbonate is a significant factor in assessing corrosivity. The charge transfer controlled reaction involving carbonic acid and carbon steel (or Fe) can be represented in terms of the concentration or partial pressure of dissolved CO₂ in the medium to arrive at a corrosion rate equation that incorporates the order of the reaction and an exponential function that approximates for Henry's reaction constant's temperature dependence. This corrosion rate equation is given as,

log (V_cor) = 5.8 - 1710/T + 0.67 log (pCO₂) ------ (1)

where

V_cor = corrosion rate in mm/yr

T = operating temperature in K

pCO₂ = partial pressure of CO₂ in bar

The corrosion rate obtained by equation (1) has typically been often seen as the maximum possible corrosion rate without accounting for iron carbonate scaling. A nomogram representing eq. (1) is given in Figure 1, which also includes a scale factor to account for the formation of protective carbonate films that lead to a reduced corrosion rate at higher temperatures.

A flow chart delineating the hierarchical reasoning structure of the predictive model is given in Figure 2. The first step in corrosivity determination is computation of the system pH, since it is the hydrogen ion concentration that drives the anodic dissolution. Further, the role of pH in promoting or mitigating CO₂-based corrosion has been extensively chronicled^22,19. For production environments, where it is the dissolved CO₂ or H₂S that contribute significantly to a suppressed pH, the pH can be determined as a function of acid gas partial pressures, bicarbonates and temperature, as shown in Figure 3. From a practical stand point, the contribution of H₂S or HCO₃ or temperature to pH determination is another way of representing effective levels of CO₂ that would have produced a given level of pH.

While it has been documented that the CO₂ corrosion mechanism is dissimilar to that of strong acids like HCl (where as CO₂ corrosion is now understood to progress through direct reduction of H₂CO₃ to HCO₃^- rather than reduction of H⁺ ions), and that carbonic acid corrosion is much more corrosive than that obtained from a strong acid such as HCl at the same pH, there is also significant agreement that lower pH levels obtained from higher acid gas presence leads to higher corrosion rates. Conversely, higher levels of pH obtained through buffering in simulated production formation water solutions have been shown to produce significantly lower corrosion rates even at higher levels of CO₂ and/or H₂S²⁴. Data about the effects of pH from another study is shown in Figure 4. Hence, it is more meaningful to determine the effective CO₂ partial pressure from the system pH. Data in Figure 3 can be represented as equations for straight lines in terms of pH and acid gas partial pressures for a given level of HCO₃ and temperature. The Predict system incorporates a numerical model to compute pH for different values of acid gas partial pressures, HCO₃ and temperature²⁵. Consequently, pH determination at a given temperature can be represented as,

pH = C1 - log (pH₂S + pCO₂) + log (HCO3)

where C1 and C2 are constants, pH₂S and pCO₂ are partial pressures in bars and HCO₃ concentration is represented in meq/l (61 mg/l).

If the temperature is higher than 100 C, there is a slight reduction in the hydrogen ion concentration as shown in Figure 3, but the change in pH can be accounted for by a change in the value of the constants in equation above. Once the system pH is determined, the effective CO₂ partial pressure can be determined from (2) as,

Log(pCO_2-eff) = (C1 - pH) / 2 ---------------- (4)

where pCO_2-eff is the effective partial pressure of CO₂ in a production system that can produce the prevalent level of hydrogen ion concentration.

The effective CO₂ partial pressure from (4) can be used in eqn. (1) to determine an initial corrosion rate for CO₂-based corrosion. The corrosion rate so obtained is modified to account for the formation of a FeCO₃ film (Fe₃O₄ at higher temperatures) whose stability varies as a function of the operating temperature. The scale correction factor shown in Figure 1 is used to determine the initial corrosion rate from the nomogram in Figure 1. It is generally estimated that this corrosion rate presents a maximum corrosion rate even though it has been reported that the rate computed by the nomogram are reached or exceeded in systems with high flow rates. It is important to recognize that this corrosion rate has to be modified to account for the effect of other critical variables in the system. Further, this rate does not indicate modality (general or localized) but rather, represents the maximum rate of attack.

As mentioned earlier, it is necessary to superposition the effects of other critical system parameters. The flow chart in Figure 2 provides the lists sequential effects that are important from a stand point of corrosivity determination. In addition to the system pH, these parameters include,

The corrosion rate predicted in the current model can be represented in terms of three broad rules that guide the computer model's decision making:

1. Effect of fundamental system variables such as CO₂, H₂S, pH, temperature, and velocity on corrosion rate.
2. Effect of parameter interactions on corrosivity, such as, influence of temperature on the carbonate or sulfide film stability. Or flow effects on corrosion products and the ensuing loss of protective films as a function of velocity, temperature, acid gases and pH.
3. Effects of system modifiers such as oil film persistence (or lack of it) or the crude type, water cut, dew point, aeration and inhibition.

Corrosion rate, thus predicted, incorporates the synergy of the effects of all the critical system variables and provides a more realistic estimation of corrosivity than what would be available with conservative theoretical models that focus on a limited number of parameters. The significance of the reasoning in predictive model stems from the fact that the decisions made synthesize different types of corrosion knowledge:

1. Theoretical models that provide effects of different parameters
2. Data from laboratory tests that provide insight on parametric correlations and trends about parametric effects
3. Experience-based heuristics that facilitate proper interpretation of data form lab and field

Data and Component Effects

The Predict system incorporates data on a number of component effects related to corrosion prediction. The Predict decision making model can be represented as shown the Figure 2. Other data in the Predict system can be classified as

CO2 Effects:

1. Corrosion nomogram
2. Effect of gas composition and temperature on corrosion rate
3. Corrosion rate as a function of temperature and CO2 pressure

H2S Effects:

1. Effect of gas composition and temperature on corrosion rate
2. Effect of H2S and temperature on corrosion rate in pure iron

Temperature Effects:

1. Corrosion rate as a function of temperature and CO2 pressure
2. Corrosion nomogram
3. Corrosion rate as a function of velocity and temperature
4. Effect of oxygen concentration as a function of temperature on corrosion
5. Effect of oxygen concentration as a function of temperature on corrosion

pH Effects:

1. In-situ pH determination for production environments
2. Corrosion rate of steel as a function of pH

Velocity Effects:

1. Corrosion rate as a function of velocity and temperature
2. Effect of gas velocity on corrosion rate

Oil type Effects:

1. Effect of acid number on crude oil wettability
2. Effect of changing crude oil type on corrosion rate as a function of water content

Oxygen Effects:

1. Effect of oxygen concentration as a function of temperature on corrosion
2. Effect of oxygen concentration as a function of temperature on corrosion

Data and Unit Conversions

PREDICT system allows utilization of both English and SI units. While the system performs an automatic conversion from English to SI and vice versa, typical conversion factors are listed in the table below for commonly utilized system parameters.

Note: M.m³ stands for millions of cubic metre and Mscf denotes Millions of standard cubic feet.

Other Conversions

Predict requires acid gas partial pressures to determine corrosion rates. If H2S and CO2 information you have is in mol percent, then partial pressure in psia is given as

Total Pressure X mol percent / 100

If acid gas content is given as ppm of total gas (from a gas analysis), then mol percent can be determined as

Mol Percent = ppm / 10000.

If dissolved acid gas content is determined from a water analysis, then the partial pressure can be determined as P = K x [C] where K = Henry's constant for the operating temperature, [C] = concentration of the component in moles/liter and P = Partial pressure in Atmospheres.