International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2199

ISSN 2229-5518

Development of a Computer Software for

Determining the Corrosion Rate of Oil
Pipelines in CO2 Environment Using
Modified NORSOK M-506 Model
IYASELE, Edgar Omondiale1, OYATOGUN Modupe Grace2
1
Department of Mechanical Engineering, Michael Okpara University of Agriculture, Umudike,
Nigeria
2
Materials Science and Engineering Department, Obafemi Awolowo University, Ile-Ife, Nigeria
*
Corresponding Author: e-mail: [email protected] Tel +234-8135359855

Abstract

The problem of Carbon dioxide (CO2) corrosion is of primary concern in Nigeria’s oil and gas

IJSER
industry. This work attempted to develop a CO2 corrosion rate prediction model, implement the

model, simulate and validate it against operating field data with the intent of developing corrosion

prediction software for the Nigerian oil and gas environment.

Existing models for CO2 corrosion prediction were accessed. Modification of one of the models; the

NORSOK M-506 model was carried out by incorporating a temperature model to it and also by

modifying the wall shear stress parameter of the model to suit the Nigerian light crude oil. The entire

model was implemented using Visual Basic to create computational software. With the developed

computational software, simulation and prediction of CO2 corrosion rate and the analysis of effects of

various parameters such as temperature, fluid velocity, pH factor, CO2 fugacity, mixture density,

mixture viscosity and wall shear stress on corrosion rate along the pipeline length were carried out

using the operating field data in the Nigerian oil and gas industry.

The mixture density and viscosity of the fluid were found to increase along the pipeline length as the

temperature decreased. These resulted in increase in the pipeline wall shear stress and consequently

increased the corrosion rate along the length of the pipeline. Other results obtained were in agreement

with similar works done in this field of study. The developed computer software also predicted

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2200
ISSN 2229-5518
corrosion rates that were in agreement with measured corrosion rates from operating field data. In

addition, the modified model had a good agreement with De Waard and Lotz semi-empirical model

and with HYDROCOR, a mechanistic model used in some of the Nigerian oil and gas industry. It had

a standard error (SE) of 0.18. And at 95% confidence level, confidence interval (CI) of the model

was between 0.70 and 1.41 mm/year mean corrosion rate predicted.

With the results obtained using the modified NORSOK M-506 model, it can be concluded that the

model can adequately replace the models presently used in the Nigerian oil and gas industry for

prediction of corrosion rates.

Keywords: Corrosion rate prediction, CO2 corrosion, corrosion, computer software, model, pipeline.

1.0 Introduction

Corrosion is defined as the destructive attack of a metal by chemical, or electrochemical, reaction

IJSER
with its environment [1]. Corrosion usually occurs because the general tendency of metals is to return

to their native form, which results in the lowering of their free energy and attainment of a more

thermodynamically stable state [2]. There are different forms of corrosion but it is rare that a

corroding structure, or component, will corrode based on only one form. Most corrosion phenomena

can be classified as either electrochemical or chemical corrosion [3]. Electrochemical corrosion

occurs when the environment is a conducting liquid; hence it is also called wet corrosion. Chemical

corrosion on the other hand occurs mainly as a result of chemical reaction of the metal with the gases

within the environment; hence it is usually referred to as dry corrosion.

The increasing cost of pipelines failures due to corrosion in the oil and gas industry has increased

interest in corrosion problems and their solutions. Corrosion failure of most facilities is often due to

the degradation of materials used for their fabrication. In design, operation, and maintenance of oil

and gas storage and production facilities, management of corrosion is critical. Materials reliability is

becoming ever more important in our society, particularly in view of the liability issues that develop

when reliability is not assured, safety is compromised, and failure occurs [1]. The problem of CO2

corrosion is the primary concern in Nigeria’s oil and gas industry as it constitutes a leading cause of

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2201
ISSN 2229-5518
pipeline failure and a main component of the operating and maintenance costs of the industries [4]. In

order to manage corrosion failures successfully, organisations have adopted a policy of “zero

failures” or “no failures.” This management policy is usually implemented by using a combination of

well-established strategies, innovative approach, and when necessary, experimental trials.

It has been reported that corrosion failure attributed to sour and sweet corrosion, which resulted from

activities of Hydrogen Sulphide (H2S) and CO2 respectively, have contributed to over 50% of all

pipelines failures [5].

The impact of corrosion in the oil and gas industry has also contributed significantly to the non-

productive time (NPT) of 20 to 30% cost from exploration to production [6]. The annual cost

associated with corrosion damage of structural components in the United States was found to be

greater than the combined annual cost of natural disasters [7].

IJSER
Pipelines transporting crude oil containing components such as CO2 and H2S have suffered internal

corrosion due to the presence of these gases dissolved in water. One of the most prevalent internal

corrosion problems in Nigerian oil and gas industry is carbon dioxide corrosion, which is also called

sweet corrosion. Hydrogen sulphide corrosion, also called sour corrosion, is not commonly

encountered in Nigeria’s oil industry.

The selection of pipe for a particular situation is dependent on the material it is transporting, the

pressure and temperature of the contents. The most commonly used material for petroleum pipelines

is carbon steel (mild steel). This may be ascribed to its strength, ductility, weldability and its

amenability to heat treatment for varying mechanical properties [8 - 11]. Carbon steel is by far the

most important alloy used in the petroleum industry and it accounts for over 98% of the construction

material, especially oil and gas transportation pipelines [12]. However, carbon steel does have more

challenging requirements such as the need for better understanding of the environment, accuracy and

reliability of the materials and corrosion performance. In both offshore and onshore oilfield

production in Nigeria, the pipelines transporting oil and gas operate with the two-phase (oil-water) or

three-phase (gas-oil-water) flow conditions. The flow system of this environment must be clearly

understood and captured in the development of corrosion prediction softwares for Nigerian oil and
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2202
ISSN 2229-5518
gas industries. In this work, the main focus is on the oil-water two-phase flow in pipelines

transporting crude oil.

The use of computational softwares by engineers for the analysis of the effect of CO2 corrosion on

pipeline materials both at the design and operational stages has been reported [13]. However,

developed computational softwares with the Nigerian content are scarce. This study attempted to

develop a computer software that will be used to determine corrosion rate of oil pipelines in CO2

environment, based on the operating parameters in the Nigerian oil and gas industry. The developed

computer software is also expected to enable the prediction of the design life of oil pipeline, the

severity level of the unmitigated corrosion rate and the mitigated corrosion rate of oil pipelines in

CO2 environment.

2.0 Research Methodology

IJSER
To enhance the achievement of the objectives of this work, the following procedures were carried out

in order to develop the CO2 corrosion prediction model:

2.1 Accessing and Modifying the NORSOK STANDARD M-506 Model

The NORSOK M-506 standard was developed with broad petroleum industry participation (Statoil,

Norsk Hydro and Saga petroleum). The model is owned by the Norwegian petroleum industry. The

model is developed to ensure adequate safety, value adding and cost effectiveness for petroleum

industry developments and operations. The NORSOK standards are normally based on recognized

international standards. They are developed according to the consensus principle generally applicable

for most standards work and according to established procedures. The NORSOK standard is limited

to a method for calculation of corrosion rates in hydrocarbon production and process systems where

the corrosive agent is CO2 [14]. The NORSOK STANDARD M-506 corrosion model for prediction

of corrosion rate, Equation 2.1, was obtained from literature [14].

𝐒 (𝟎.𝟏𝟒𝟔+𝟎.𝟎𝟑𝟐𝟒𝐥𝐨𝐠(ƒ𝑪𝑶𝟐 ))
𝑪𝑹𝒕 = 𝑲𝒕 × (ƒ𝑪𝑶𝟐 )𝟎.𝟔𝟐 × � � × ƒ(𝐩𝐇)t (2.1)
𝟏𝟗
CRt = corrosion rate at temperature‘t’ in mm/year, Kt = equilibrium constant at temperature t, ƒco2 =

fugacity of CO2 in bar, S = wall shear stress in Pascal (Pa) and ƒ(pH)t = pH factor at temperature t).

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2203
ISSN 2229-5518
2.2 Development of the Temperature Distribution Model

The equation used to obtain the temperature of fluid pumping over the pipeline section of length (L)

in cross sections (x), Equation (2.2), was developed using the conservation of energy equation for

steady flow of a fluid in a pipe.

𝐗
𝐓𝐞(𝐋)– 𝐓𝐬 𝐋
𝐓𝐞(𝐱) = 𝐓𝐬 + �𝐓𝐢 – 𝐓𝐬� × � � (2.2)
�𝐓𝐢 – 𝐓𝐬�
Where Te = exit mean fluid temperature, Ti = inlet mean fluid temperature, Ts = constant.

2.3 Predicting the Temperature Dependent Constant (Kt)

The constants for the temperature t, (Kt) were generated by NORSOK STANDARD M-506 for

temperatures between 20oC and 150oC. These values were used in obtaining simulated temperatures

by a linear extrapolation at the temperature above and below the desired temperature.

2.4 Predicting the pH Factor

IJSER
With the predicted temperature, and a given pH value of the fluid, the pH factor of that particular

temperature was predicted. pH function at different temperatures used in this work is as given by

NORSOK STANDARD M-506 (2005) [14].

2.5 Development of the CO2 Fugacity Model

The equation for calculating the fugacity of CO2 (ƒCO2) was derived by multiplying the fugacity

constant with the partial pressure of a gas, as shown in equation (2.3). This was utilized to calculate

the fugacity of CO2 for the system.

ƒ𝐂𝐎 = 𝐚 × 𝐏𝑪𝑶𝟐 (2.3)

𝟐
Where ƒCO2= fugacity of CO2, a = fugacity coefficient, Pco2= CO2 partial pressure

NORSOK STANDARD M-506 (2005) [14] derived the fugacity coefficient, Equation (2.4) used to

compute the fugacity coefficient as:

𝟏.𝟒 𝟏.𝟒
�𝐏 × �𝟎.𝟎𝟎𝟑𝟏 – ��
𝐚 = 𝟏𝟎 𝐓 𝐟𝐨𝐫 𝐏 ≤ 𝟐𝟓𝟎 𝐛𝐚𝐫, 𝐚 = 𝟏𝟎(𝟐𝟓𝟎 × (𝟎.𝟎𝟎𝟑𝟏 – ))
(2.4)
𝐓 𝐟𝐨𝐫 𝐏 > 250 𝑏𝑎𝑟
O
(Where T is the temperature ( K) calculated using equation (2.2), while P, is the total system

pressure in bar). The program was built to convert the Celsius unit to Kelvin.

2.6 Development of the Wall Shear Stress Model

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2204
ISSN 2229-5518
Taitel et al., [15] gave the Blasius type equation for calculating oil layer, mixed layer, water layer and

the interfacial wall shear stresses as shown below:

�ƒ𝐰 ×ρ𝐰×(𝐔𝐬𝐰 )𝟐 �
𝐒 = τ𝐰 = (2.5)
𝟐
�ƒ𝐦 ×ρ𝐦×(𝐔𝐬𝐭 )𝟐 �
𝐒 = τ𝐦 = (𝟐. 𝟔)
𝟐
�ƒ𝐨 ×ρ𝐨×(𝐔𝐬𝐨 )𝟐 �
𝐒 = τ𝐨 = (2.7)
𝟐
Equations (2.5), (2.6) and (2.7) above are wall shear stresses of water, mixed and oil layers

respectively. Subscripts w, m and o means water, mixed and oil respectively..

��ƒ𝐢𝟏 ×ρ𝐰 ×(𝐔𝐬𝐭 −𝐔𝐬𝐰 )�×|𝐔𝐬𝐭 −𝐔𝐬𝐰 |�

𝐒 = τ𝐢𝟏 = (2.8)
𝟐
{(ƒ𝐢𝟐 ×ρ𝐨 ×(𝐔𝐬𝐨 −𝐔𝐬𝐭 ))×|𝐔𝐬𝐨 −𝐔𝐬𝐭 |}
𝐒 = τ𝐢𝟐 = (2.9)
𝟐
(Where subscripts i1 and i2 are interfaces of water layer-mixed layer and mixed layer-oil layer

IJSER
respectively).

Mean wall shear stress was then calculated by adding all the wall shear stresses: Equations (2.6) to

(2.9) as given below;

τ = τw+ τm+ τo+ τi1+ τi2 (2.10)

Where τ=wall shear stress (Pa), ƒ= friction factor, ρ= density of fluid, Ust = oil-water superficial

velocity, Uso = superficial oil velocity and Usw = superficial water velocity.

2.6.1 Friction factor

In order to model the friction factor, f, Equation 2.11 derived by using the Fanning friction factor,

which for turbulent flow can be approximated [16] was adopted.

The friction factor calculated is valid for Reynolds numbers ranging from 3000 to 3,000,000.

(𝟎.𝟎𝟒𝟔)
ƒ = (2.11)
𝐑𝐞𝟎.𝟐
Where Re, the Reynolds number, is obtained using equation 2.12 below;
(ρ 𝐔 𝐃)
𝐑𝐞 = (2.12)
µ
(µ is the dynamic viscosity (Ns/m2), D is the pipe diameter (m) and U is fluid velocity).

The friction factors for the different layers, water, mixed and oil layer were obtained using the

Renolds numbers for the different phases:

(𝟎. 𝟎𝟒𝟔)
ƒ𝒎 = (𝟐. 𝟏𝟑)
(𝐑𝐞𝐦 )𝟎.𝟐
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2205
ISSN 2229-5518
(𝟎.𝟎𝟒𝟔)
ƒ𝒘 = (2.14)
(𝐑𝐞𝐰 )𝟎.𝟐
(𝟎.𝟎𝟒𝟔)
ƒ𝒐 = (2.15)
(𝐑𝐞𝟎 )𝟎.𝟐
The interfacial friction factors ƒi1 and ƒi2 are set as equal to the wall friction factors of the faster
phases as shown below [20]:
ƒ𝐢𝟏 = ƒ𝐦 and ƒ𝐢𝟐 = ƒ𝐨
According to equation (2.12), the Reynolds numbers defined for each phase as a function of the
system temperature are;
�ρ𝐰 × 𝐔𝐬𝐰 × 𝐃�
𝐑𝐞w = (2.16)
µ𝐰
�ρ𝐦 × 𝐔𝐬𝐦 × 𝐃�
𝐑𝐞m = (2.17)
µ𝐦
�ρ𝐨 × 𝐔𝐬𝐨 × 𝐃�
𝐑𝐞o = (2.18)
µ𝐨
2.6.2 Density of fluid

Water density as a function of temperature was calculated using the Equation 2.19, which was

derived by Gray et al., [17].

w = 𝟏𝟏𝟓𝟐. 𝟑 – 𝟎. 𝟓𝟏𝟏𝟔 × 𝐓 (2.19)

IJSER
Where T is temperature in degrees Celsius and density is measured in (Kg/m3) Density of oil was

computed using equation (2.20), which was obtained from [18];

ρo = ρ15 × [𝟏 − 𝟎. 𝟎𝟎𝟎𝟕(𝐓 − 𝟏𝟓)] (2.20)

(Where T is temperature in degrees Celsius and ρ15 is oil density at 15oC which was taken as
3
880kg/m ).

The mixture density was calculated as the weighted average of the two phases based on the hold-up

of the dispersed phase, Equation 2.21, [19].

ρm= (𝟏 − εd) ρc + εd d (2.21)

Where ε is water cut or hold-up and subscripts c and d are used for continuous and dispersed phase

respectively.

2.6.3 Viscosity of fluid

To calculate water viscosity as a function of temperature, we made use of the expression given by the

[20]:
𝟐
�𝟏.𝟑𝟐𝟕𝟐�𝟐𝟎 –𝐭�– 𝟎.𝟎𝟎𝟏𝟎𝟓𝟑�𝟐𝟎 – 𝐭� �
{ }
µ = µref.× 𝟏𝟎 𝐓 + 𝟏𝟎𝟓 (2.22)

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2206
ISSN 2229-5518
Where µref= reference dynamic viscosity of water at 20oC, whose value is often given as 0.001002

Ns/m2 [21], T is absolute temperature (K) and t is temperature in degree Celsius.

For oil viscosity, the modified Ng and Egbogah’s correlation (1983) [22] for light crude oil in

Nigeria was used as shown in equation (2.23):

(𝟏.𝟔𝟕𝟎𝟖𝟑 – ( 𝟎.𝟎𝟏𝟕𝟔𝟐𝟖 × 𝐀𝐏𝐈) – 𝟎.𝟔𝟏𝟑𝟎𝟒 × 𝐋𝐨𝐠 (𝐓))

od = 𝟏𝟎𝟏𝟎 (2.23)

T is temperature (oF) and API is oil gravity used to describe oil composition and measured in oAPI.

Oil viscosity measured in Centipoise (cP)

To determine the specific gravity of the oil, we made use of the Equation (2.24) [21] given as:

(𝟏𝟒𝟏.𝟓)
𝐒𝐆oil = (2.24)
(𝟏𝟑𝟏.𝟓+ 𝐀𝐏𝐈)
𝟏𝟒𝟏.𝟓
𝐖𝐡𝐞𝐫𝐞 𝐀𝐏𝐈 = ( – 𝟏𝟑𝟏. 𝟓) (2.25)
𝐒𝐆
To calculate the liquid mixture viscosity, Lyons [23] derived the expression shown in Equation

IJSER
(2.26). This was used to calculate the liquid mixture viscosity in this work.
𝟏 𝐖𝐎𝐑
m= µo[(𝟏+𝐖𝐎𝐑)] + µw( ) (2.26)
{𝟏+𝐖𝐎𝐑}

Where WOR = water-oil ratio, i.e. water cut and the mixed viscosity is measured in (Ns/m2)

2.6.4 Superficial velocity of fluid

The superficial fluid velocity (m/s) was calculated by dividing the flow rate with the pipe cross

section area, A, both for the oil, Uso, and water, Usw as shown equations (2.28) and (2.29). The total

oil–water superficial velocity, Ust is the combined velocity of oil and water used. It was derived by

Cai [19] and shown in equation (2.27).

𝐔st = 𝐔so + 𝐔sw (2.27)

𝐐𝐰
𝐔sw= , (2.28)
𝑨
𝐐𝐨
𝐔so= (2.29)
𝑨
Where Qw = total input volume flow rate of water phase and QO = total input volume flow rate of oil

phase, which were obtained using Equations 2.30 and 2.31 below:

𝐐w = 𝐐wL+ εm 𝐐m (2.30)
QO = 𝐐OL+ (𝟏 − εm) 𝐐m (2.31)
Where εm = water cut in the oil-water mixed layer. εm at the point of phase inversion εinvert was

calculated using the expression as given by Arirachakaran [24],

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2207
ISSN 2229-5518

m= εinvert = 𝟎. 𝟓 – 𝟎. 𝟏𝟏𝟎𝟖𝐥𝐨𝐠𝟏𝟎 (𝟏𝟎𝟑 × µo) (2.32)

QwL = pure water layer flow rate, QOL = pure oil layer flow rate, Qm = oil-water mixed layer flow rate

and µ is oil viscosity. In this work, we assumed Qm to be given by the relation:

𝐐m = 𝐐wl+ 𝐐ol (2.33)

2.7 Predicting the Uninhibited CO2 Corrosion Rate

With Equation (2.1), we predicted the uninhibited CO2 corrosion rate (CRt) along the oil pipeline by

inputting the values predicted by models of Kt, ƒCO2, S, and ƒ(pH)t.

2.7.1 Predicting the severity level of the uninhibited corrosion rate

This was done using the categorized severity levels as defined by (Guidelines for Prediction of CO2

Corrosion in Oil and Gas Production Systems, 2009) [25].

2.7.2 Predicting the life expectancy of the pipeline material

IJSER
�𝐗 – 𝐘�
𝐋𝐢𝐟𝐞𝐭𝐢𝐦𝐞 𝐨𝐟 𝐩𝐢𝐩𝐞 𝐦𝐚𝐭𝐞𝐫𝐢𝐚𝐥 = (2.34)
𝐂𝐑
The life expectancy of the pipeline, measured in years, CR is the calculated corrosion rate in

mm/year, where X = pipe material thickness before degradation in mm and Y= minimum allowable

pipe thickness during degradation before replacement in mm.

2.7.3 Predicting the mitigated CO2 corrosion rate

The mitigated corrosion rate (CR) was calculated using equation (2.35) as given by Gartland and

Johnsen [26]. This was derived based on the predicted unmitigated corrosion rate and the inhibitor

efficiency.

�𝟏𝟎𝟎 – 𝐈𝐄�
𝐂𝐑 = 𝐂𝐑unmit × (2.35)
𝟏𝟎𝟎
CRunmit= unmitigated corrosion rate in mm/year, IE = inhibitor efficiency in percentage.

2.8 Implementation, Simulation and Validation of the Model

Implementation of the models was done using Visual Basic, and a computational software

application was created to run under Microsoft windows environment. Figure 2.1 shows the user’s

interface of the computational CO2 corrosion rate prediction program; Figure 2.2 shows the CO2

corrosion rate calculation flowchart;

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2208
ISSN 2229-5518
The simulation input parameters were obtained from the oil company operating the Oil Mining

License (OML 61) and were used as input data to simulate corrosion rate along the pipeline

and to also analyze the effects of different parameters. The model results were validated against

data taken from Nigerian oil and gas industry for pipelines transporting oil and water. To further

validate this model, a semi-empirical model for corrosion rate with respect to velocity in the absence

of surface scale with a parallel resistance was used [27].

The result of their experimental analysis is summarized in the following expression as;

𝟏 𝟏 𝟏
= + (2.36)
𝐕𝐜𝐫 𝐕𝐫 𝐕𝐦
Where Vcr = corrosion rate, Vr = flow independent contribution denoting the reaction rate

Vm = flow dependent contribution denoting the mass transfer rate [27]

𝟏𝟏𝟏𝟗
Log(Vr)= 𝟒. 𝟗𝟑 − + 𝟎. 𝟓𝟖𝐥𝐨𝐠 (𝑷𝑪𝑶𝟐 ) (2.37)
𝑻+𝟐𝟕𝟑

IJSER
o
Where T = Temperature ( C), Pco2= C02 Partial Pressure (MPa) and
�𝑼𝟎.𝟖 �
𝑽m= 𝟐. 𝟒𝟓 × × 𝑷𝑪𝑶𝟐 (2.38)
(𝑫𝒉𝟎.𝟖 )
Where U = liquid flow rate (m/s) and Dh = hydraulic diameter of the pipe.

Figure 2.1 The User’s Interface of the Computational CO2 Corrosion Rate Prediction Software

IJSER © 2018
http://www.ijser.org
 International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2209
ISSN 2229-5518 START

Input pH, Inlet Temp., Surrounding Temp., Heat Transfer

Coefficient, Heat capacity, Diameter, Density, Velocity,
Total Pipe Length, Distance x along Pipe, pipe thickness,
Pressure, PCO2, API, Water Flowrate, Oil Flowrate,
Inhibitor Efficiency.

Water cut =Qw / (Qw + Qo) Mass Flow Rate = ρVA

Exit temperature at total pipe length and distance x along pipe length, (Te(x)) = Ts +Ti – Ts {(Te(L) – Ts)/(Ti Ts)}x/L

a = 10P * (0.0031 – 1.4/T) YES Pressure <= 250 bar NO a = 10250 * (0.0031 – 1.4/T)

CO2Fugacity = a * Pco2, FluidDensity = 1152.3 – 0.5116 * T, OilDensity = ρ15[1− 0.0007(T−15)], MixtureDensity = (1 - εd) ρc +
εdρd, MixtureViscosity = µo [1/(1+WOR)] + µw (WOR/[1+WOR]), SuperficialMixture Velocity = Ust = Uso + Usw,
MixtureReynoldsNumber = (ρm Ust D) / µm, MixtureFrictionFactor = (0.046) / Rem0.2 , WallShearStress = τw + τm + τo + τi1 + τi2, C
R = Kt (ƒco2)0.62 × (S/19)0.146+0.0324log(ƒco2) × ƒ(pH)t, InhibitedCR = CRunmit * (100 – IE) / 100 , LifetimeofPipe = (X – Y) / CR

Mass Flow Rate, Water Cut, Exit Temp., kt, pH Factor, CO2fugacity,
Fluid Density, Oil Density, Mixture Density, Mixture Viscosity,
Superficial Mixture Velocity, Mixture Reynolds Number, Mixture
Friction Factor, Wall Shear Stress, CR(uninhibited), CR(Inhibited), Severity

IJSER
levels, Life Time of Pipe

STOP
Figure 2.2 The CO2 Corrosion Rate Calculation Flowchart

3.0 Results and Discussions

The correlation results of the corrosion rate in the pipeline as a dependent variable and temperature

along the pipeline distance are shown in figures 3.1 to 3.3.

66 PIPE DISTANCE VS TEMPERATURE @ FLUID

70 VELOCITY 0.31m/s
56
60
47
50
Temperature (oC)

40
40 34
29
30 24
20 20 20 20
20

10

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Pipe Distance (m)

Fig 3.1 Pipe Distance Versus Temperature

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2210
ISSN 2229-5518

CORROSION RATE VS DISTANCE ALONG PIPE LENGTH

3 2.692
2.456
2.275
Corrosion Rate (mm/yr)
2.5
1.94
2 1.67
1.51 1.395
1.5 1.18 1.18 1.18 1.18

0.5

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Distance (m)

Fig 3.2 Corrosion Rate along the Pipeline Length

CORROSION RATE VS TEMPERATURE, OTHER VARIABLES NOT

3 2.692
CONSTANT ALONG THE PIPE
2.456
2.5 2.275
Corrosion Rate (mm/yr)

1.94

IJSER
2 1.67
1.51
1.395
1.5 1.18 1.18 1.18 1.18

0.5

0
66 56 47 40 34 29 24 20 20 20 20
Temperature (oC)

Fig 3.3 Corrosion Rate versus Temperature along the

Pipeline Length
It was observed that the temperature gradually decreased as the oil flows through the pipeline, see

Figure 3.1. This can be ascribed to the gradual decrease in temperature from inlet temperature at

distance (0 metres) due to heat transfer from the heated oil to the surrounding. In isothermal pipelines

(where no intermediate heating stations are installed to warm the oil) and in pipelines without heat

insulation, the temperature will eventually decline to the surrounding temperature at some kilometers

after the inlet point, depending on many factors such as the surrounding temperature, the overall heat

transfer coefficient, velocity, and fluid heat capacity [28].

From Figure 3.2, it was noticed that there was a change of corrosion rate as the fluid flowed via the

total pipeline length. Also, Figure 3.3 showed a similar result as the temperature was varied with

corrosion rate. It was observed that from Figure 3.2 and Figure 3.3, the corrosion rate increased to a
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2211
ISSN 2229-5518
o
particular temperature and distance (56 C and 1,000 metres in this case) and decreased as the fluid

flows over the pipeline distance until it remained constant at a particular temperature and distance (20
o
C and 7,000 metres in this case). This is in agreement with the findings of Nesic (2007) who

established that the peak in the corrosion rate is usually seen between 60 oC and 80 oC, depending on

water chemistry and flow conditions [29].

The modeled result for pipeline corrosion rate was compared with HYDROCOR at different

operating conditions; with OML 61 at different temperatures and DE WAARD AND LOTZ model

also at different temperatures. The results are shown in figures 3.4 to 3.6.

DEVELOPED SOFTWARE VS HYDROCOR

0.9
0.8 DEVELOPED SOFTWARE
Corrosion Rate (mm/yr)

0.7 HYDROCOR

IJSER
0.6
0.5
0.4
0.3
0.2
0.1
0
1 2 3
Operating Condition

Fig 3.4 Developed Software with HYDROCOR at Different Operating Conditions

DEVELOPED SOFTWARE VS OML 61
DEVELOPED SOFTWARE
OML 61
1.6 1.436 1.453
1.4
Corrosion Rate (mm/yr)

1.2
0.96 0.939
1
0.8 0.727 0.729
0.6 0.494 0.482
0.409
0.4 0.259
0.2
0
46 59 61 64 70
Temperature (oC)

Fig 3.5 Developed Software With OML 61 at Different Temperatures

IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2212
ISSN 2229-5518

DEVELOPED SOFTWARE VS DE WAARD & LOTZ

DEVELOPED SOFTWARE
1.49 DE WAARD AND LOTZ
1.6 1.436 1.453 1.44
1.4 1.28 1.31

Corrosion Rate (mm/yr)

1.175
1.2
0.952 0.96 0.939
1
0.8
0.6 0.482
0.34
0.4
0.2
0
46 59 60 61 64 70
Temperature (oC)

Fig 3.6 Developed Software with DE WAARD AND LOTZ Model at Different Temperatures

Figure 3.4 showed the comparison of the developed software with SHELL’S HYDROCOR. The

results are in good agreement. The developed software is an empirical model while HYDROCOR is a

IJSER
mechanistic model. In practice, it is a well known fact that the mechanistic model HYDROCOR is

very conservative when it comes to predicting corrosion rate. That is, the model under predicts when

compared to other models especially those that do not put into much consideration the physics behind

the corrosion just as mechanistic models do.

To ascertain the statistical accuracy of the predicted values, standard error (SE) calculation was

performed. HYDROCOR had a SE of 0.08 while the developed software had a SE of 0.2. Also, in

order to ascertain the reliability of the predicted values, confidence interval (CI) calculation was

carried out. At confidence level of 90%, HYDROCOR had a confidence interval between 0.04

mm/year and 0.32mm/year. This is to say that at 90% confidence level, it can be said, we are 90%

confident the mean corrosion rate is between 0.04 and 0.32mm/year. At the same confidence level of

90%, the developed software had a confidence interval between 0.04 and 0.73 mm/year. This is to

say also that at 90% confidence level, it can be said, we are 90% confident the mean corrosion rate

predicted is between 0.04 and 0.73mm/year.

Furthermore, the developed software prediction was compared with operating field data obtained

from a Nigerian oil and gas company (OML 61). The results obtained are shown in Figure 3.5.To

ascertain the accuracy of the model prediction, the average percent error (APE) and average absolute
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2213
ISSN 2229-5518
percent error (AAPE) of the data source used were calculated. The APE is used to quantify the

degree of over prediction or under prediction of the experimental data. Positive values indicate over

prediction while negative values indicate under prediction. AAPE is used to evaluate the prediction

capability of the model. The lower the value of AAPE; the better the precision of the calculated

values that can be achieved from the model. The APE for the developed software was 19.74%,

meaning the model over predicted compared to the operating field data. Often times, predicted values

and measured values are never exactly the same. The developed software also had an AAPE of

19.74%. This value cannot be said to be accurate or not since the range of percentage error where it is

regarded as accurate or not accurate is a function of application.

To measure the statistical accuracy of the predicted values, SE was calculated. The developed

software had a SE of 0.18. To determine the reliability of the predicted values, confidence interval

IJSER
(CI) was calculated. At 95% confidence level, the developed software had a CI between 0.70 and

1.41 mm/year. This means that we are 95% confident that the mean corrosion rate predicted is

between 0.70 and 1.41 mm/year.

For the sake of more comparison using the same operating field data, the De Waard and Lotz semi

empirical model was employed, see Figure 3.6. The Waard and Lotz semi empirical model had an

APE of 19.34%, AAPE of 19.34%, SE of 0.21 and confidence interval (CI) at 95% confidence level

of between 0.74 and 1.55 mm/year. Comparing the developed software with De Waard and Lotz

model, we had a good agreement as depicted by Figure 3.6. From the statistical analysis results

mentioned earlier for both models, we could see that both models over predicted having a positive

APE value; statistical accuracy of predicted values of this model was better when compared with De

Waard and Lotz; and the reliability of the predicted values at 95% confidence level of both models

were between 0.7 and 1.41mm/year for this model and 0.74 and 1.55mm/year for De Waard and Lotz

model.

4.0 Conclusion

A computer software for determining the corrosion rate of oil pipelines in CO2 environment using

modified NORSOK M-506 model has been developed. The NORSOK M-506 model has been
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2214
ISSN 2229-5518
modified to include a temperature distribution model that predicts the temperature at any point along

the oil pipeline. The wall shear stress parameter which is a function of flow related parameters

(viscosity, density, friction and velocity) of the NORSOK M-506 model has also been modified to

suit the light crude oil of the Nigerian oil and gas industry.

The entire model has been successfully implemented using an enhance-able; user friendly Visual

BASIC to create a computational software that can run under any Microsoft windows.

Simulation and validation of the model was carried out. The simulation and validation results showed

an acceptable agreement. The validation done by comparing the model with SHELL’S HYDROCOR

and De Waard and Lotz models was satisfactory, as this model recorded an acceptable agreement

with the mechanistic and semi empirical models of HYDROCOR and De Waard and Lotz. The

developed computational software is user friendly and has a computation time of less than one

IJSER
minute for the user to input data and see the output.

References

[1] Revie, W.R. and Uhlig, H.H. (2008). Corrosion and Corrosion Control: An Introduction to

Corrosion Science and Engineering, Fourth Edition. A John Wiley and Sons, Inc, Publication,

New Jersey.

[2] Mantell, C.L. (1958) Engineering Materials Hand Book. McGraw Hill Company, New York

(First Edition), pp.37.

[3] Rajput, R.K. (2010) A Textbook of Material Science and Engineering (S.I Units). S.K.

KATARIA and Sons, 6, Guru Nanak Market, Nai Sarak, Delhi, pp. 854 – 866.

[4] Oyatogun, G.M. Adeoye, M.O. Olorunniwo, O.E. and Umoru, L.E. (2011), “Failure Analysis of a

Ruptured Trunkline,” Journal of Failure Analysis and Prevention, Vol. 11, No. 5, pp 522- 529

[5] Canadian Association of Petroleum Producers (CAPP) (2012), http://www.capp.ca/forecast.

[6] Nicholson, R. Feblowitz, J. Madden, C. and Bigliani, R. (2010) “The Role of Predictive Analytics

in Asset Optimization for the Oil and Gas Industry – White Paper” http://www.tessella.com/wp-

content/uploads/2008/02/IDCWP31SA4Web.pdf
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2215
ISSN 2229-5518
[7] Thacker, H. B. Glenn, M. L. James, F. D. Elizabeth, T. Fengmei, S. Carl, F.P. Kent, E. C.

and Richard, A. P. (2010) Corrosion Control in Oil and Gas Pipelines, South west Research

Institute, San Antonio, TX. Vol. 237 No. 3, ISSN 1547 -7029

[8] Smith, W.F. and Hashemi, J. (2006) “Foundations of Material Science and Engineering”

Fourth Edition, McGraw Hill, New York.

[9] AbdulHameed, A. (2005) “Evaluation of Quality of Steels Produced Locally” Unpublished

B.Eng. Thesis, Department of Mechanical Engineering,University of Ilorin, Ilorin, Nigeria.

[10] Bolton, W. (1994) “Engineering Materials Technology” 2nd Edition, B.H Newnes Ltd,

Oxford, London, England.

[11] Davies, D. J. and Oelmann, L. A. (1983) “The Structure, Properties and Heat Treatment of

Metals” First Edition, Ritman Publishing, New Zealand Ltd, Wellington, England.

IJSER
[12] Bodude, M.A., Adeosun, S.O. and Ayoola, W.A. (2012) Comparative Studies on Mechanical

and Corrosion Characteristics of API 5LX60 Steel and RST 37-2 Steel, Journal of Emerging

Trends in Engineering and Applied Sciences (JETEAS) 3(1): 137-143.

[13] Guidelines for Prediction of CO2 Corrosion in Oil and Gas Production Systems (2009), Institute

for Energy Technology (IFE). Report No. IFE/KR/E -2009/003, ISSN 0333 – 2039, Norway.

[14] CO2 Corrosion Rate Calculation Model, Rev. 2 (2005) NORSOK standard No.

M-506,http://www.standard.no/en/Sectors/Petroleum/NORSOKStandard

Categories/MMaterial/M-5061, Standards Norway.

[15] Taitel, Y. Barnea, D. and Brill, J.P. (1995) Stratified Three Phase Flow in Pipes. International

Journal of Multiphase Flow, 21, 53-60.

[16] Knudsen, J.G. and Katz, D.L.V. (1954) Fluid Dynamics and Heat Transfer. University of

Michigan.

[17] Gray, L. G. S. Anderson, B. G. Danysh, M. J. and Tremaine, P. R. (1990) ʺEffect of pH and

Temperature on the Mechanism of Carbon Steel Corrosion by Aqueous Carbon Dioxideʺ,

Corrosion/90, paper no. 40, Houston, TX: NACE International.

[18] Knezevic, D. and Savic, V. (2006) Mathematical Modeling of Changing of Dynamic Viscosity,
IJSER © 2018
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 9, Issue 5, May-2018 2216
ISSN 2229-5518
As a Function of Temperature and Pressure, of Mineral Oils for Hydraulic Systems.

FACTAUNIVERSITATIS Series, Mechanical Engineering, Vol 4, No. 1, pp. 27 – 34.

[19] Cai, J. Chong, L. Xuanping, T. Francais, A. Sonja, R. and Nesic, S. (2012) Experimental

Study of Water Wetting in Oil-Water two Phase Flow – Horizontal Flow of Model Oil.

Institute for Corrosion and Multiphase Technology, Ohio University, USA.

[20] Handbook of Chemistry and Physics (1985) 66th ed. Boca Raton, Fl: CRC Press Inc., p. F-37.

[21] Wikipedia, 9th January 2014. http://www.wikipedia.com

[22] Ng, J.T.H. and Egbogah, E.O. (1983) “An Improved Temperature-Viscosity Correlation for

Crude Oil Systems,” paper 83-34-32 presented at the 34th Annual Technical Meeting of the

Petroleum Society of CIM, Banff.

[23] William, C. L. (ed) (1996) Standard Handbook of Petroleum and Natural Gas Engineering

IJSER
(Vol 1), Gulf Publishing Company, Houston Texas.

[24] Arirachakaran, S. Oglesby, K.D. Malinowsky, M.S. Shoham, O. and Brill J.P. (1989) An

Analysis of Oil/Water Flow Phenomena in Horizontal Pipes, SPE Paper 18836,

SPEProduction Operations Symposium, pp. 155-167, Oklahoma.

[25] Guidelines for Prediction of CO2 Corrosion in Oil and Gas Production Systems (2009), Institute

for Energy Technology (IFE). ReportNo. IFE/KR/E -2009/003, ISSN 0333 – 2039, Norway.

[26] Gartland, P. O. and Johnsen, R. (2003) Application of Internal Corrosion Modeling in the Risk

Assessment of Pipelines, CorrOcean ASA, Norway, paper No. 03179

[27] De Waard, C. Lotz, U. and Dugstad, A. (1995) "Influence of Liquid Flow Velocity on CO2

Corrosion: A Semi-Empirical Model", CORROSION/95, Paper No. 128, NACE International.

[28] Mohyaldin, M.E. Elkhatib, N. and Ismail, M.C. (2011) Coupling NORSOK CO2 Corrosion

Prediction Model with Pipelines Thermal/Hydraulic Models to Simulate CO2 Corrosion along

Pipelines. Journal of Engineering Science and Technology, Vol. 6, No. 6, 709 – 719.

[29] Nesic, S. (2007) Key Issues Related to Modelling of Internal Corrosion of Oil and Gas Pipelines

– A Review. Elsevier Journal on Corrosion Science, Corrosion Science 49, 4308 – 4338.

IJSER © 2018
http://www.ijser.org