Methods of Testing and ure Corrosionan Corrosion by Hoi Gases

osion (Proceedings of /orkshop)on Microbial Corrosion

Materials Requirements jteels for H2S- n Oil and Gas

? on Corrosion :>i Oil and

stant Alloys for Oil andan Generalthods for H^S Service/ on Corrosion in Oil and

in Concrete:

/ on Corrosion of

itrol of Stainless Steels encesies o h Marine Corrosion and

norganic Coatings for ^search and Experiences

noreanic Coatings for. and Experiences

crobially Induced

i ,iiui iltc EFC Working

U M a

European Federation of Corrosion Publications

NUMBER 23

A Working Party Report on

C 0 2 Corrosion Control in Oil and Gas Production

Design ConsiderationsEdited by

M. B. K erm an i & L. M . S m ith

Published for the European Federation of Corrosion by The Institute of Materials

THE INSTITUTE OF MATERIALS 1997

Contents

Series Introduction .........................................................................................................................vii

P reface .......................................... ................ .......................................................................................ix

Acknozuledgements................................ .............. ........ ................................ ......................... x

1 Introduction.......................................................... .................................................................1

2 S co p e ..................................... ...... ......................................................................................... 3

3 The Mechanism of CO2 Corrosion.......... ........ ...........................................................4

4 Types of CO2 Corrosion D am age.................. ................................................................ 6

4.1. Localised Corrosion of Carbon S te e l...................................................................... 6

4.2. Localised Corrosion of Carbon Steel W eld s......................................................... 7

5 Key Parameters Affecting Corrosion.............................................................................9

5.1. Water W etting............................................ ....................................................................9

5.1.1. Water Characteristics.......................................................................................10

5.1.2. Hydrocarbon Characteristics........................................................................ 10

5.1.3. Top-of-the-Line W etting............... ................................................................. 11

5.2. Partial Pressure and Fugacity of C 0 2 ..................................................................12

5.3. Temperature...............................................................................................................12

5.4. p H ................................................................................................................................. 14

5.5. Carbonate Scale...................................... .................... ..............................................15

5.6. The Effect Of H2S ...................................................................................................... 15

5.7. Wax Effect................................................................................................................... 16

6 Prediction of the Severity of CO 2 C orrosion .........................................................18

6.1. C 0 2 Corrosion Prediction Models For Carbon S tee l..................................... 19

7 CO 2 Corrosion C ontrol.......................................................... ........................................ 24

7.1. Micro-alloying of Carbon and Low Alloy S te e ls ............................................ 24

7.1.1. Effect of Chromium.............................................. ............................................24

7.1.2. Effect of C arbon .................................................................................................25

7.1.3. Effect of Other Alloying Elem ents...............................................................25

7.2. Effect of Glycol and M ethanol.............................................................................. 26

in Contents

7.3. pH Control........................ ............................................................................................27

7.3.1. The Role of p H .................................................................................................. 27

7.3.2. Wet Gas Transportation L in es........................................................... .......... 27

7.3.3. Different Chemicals and Their M echanism s.............................................. 27

7.3.4. pH M onitoring..................................................................................................... 28

7.4. Corrosion Inhibition...................................................................................................28

7.4.1. Inhibitor M echanism .......................................................................................... 29

7.4.2. Inhibitor Efficiency and Inhibitor Perform ance...................................... 30

7.4.3. Inhibitor Partitioning and Persistency ......................................................... 31

7.4.4. Commercial Inhibitor Packages................................................ ......................34

7.4.5. Inhibitor Compatibility..................................................................................... 34

7.4.6. Inhibitor Deploym ent........................................................................................35

7.4.7. Inhibitor Distribution in Multiphase Pipelines..........................................36

7.4.8. Effect of Flow on Inhibition.............................................................................36

8 Corrosion Allowance D eterm ination ..................... .......................................... ......37

8.1. Design Corrosion Allowance..................................................................................38

8.1.1. Design Corrosion R ate ...................................................................................... 38

8.1.2. Design Corrosion Allowance Assessment................................................. 38

9 Design C onsiderations.................................................................. ......... .................... 41

9.1 Well Com pletions........................................................................................................41

9.1.1. Corrosion D esign................................................................................................42

9.1.2. Corrosion Monitoring ....................................................................................... 43

9.2. Production Facilities.................................................................................................. 44

9.2.1. Corrosion Design......................... ...................................................................... 44

9.2.2. Multiphase Fluid Behaviour...........................................................................46

9.2.3. Corrosion M onitoring....................................................................................... 47

9.3 Gas Reinjection.............................................................................................................49

9.3.1. General Requirements for Gas Reinjection.......... ........... ...........................49

9.3.2. Onshore Delivery Lines....................................................................................49

9.3.3. Offshore Delivery Lines.................................................................................... 50

9.3.4. Injection Wrells And Gas Lift Annuli ............................................................ 50

References......................................................................................................................................... 51

organisemateriathrough

Te. V. in < Materia' Ad mini: from all activitie: corrosio of activi

TIexample technica often pu

Inpublishe tion are ,c Institute

I Iwith efft

A. D. MLEFC Sen The Inst

European Federation of Corrosion Publications Series Introduction

The EFC, incorporated in Belgium, was founded in 1955 with the purpose of promoting European co-operation in the fields of research into corrosion and corrosion prevention.

Membership is based upon participation by corrosion societies and committees in technical Working Parties. Member societies appoint delegates to Working Parties, whose membership is expanded by personal corresponding membership.

The activities of the Working Parties cover corrosion topics associated with inhibition, education, reinforcement in concrete, microbial effects, hot gases and combustion products, environment sensitive fracture, marine environments, surface science, physico-chemical methods of measurement, the nuclear industry, computer based information systems, the oil and gas industry, the petrochemical industry and coatings. Working Parties on other topics are established as required.

The Working Parties function in various ways, e.g. by preparing reports, organising symposia, conducting intensive courses and producing instructional material, including films. The activities of the Working Parties are co-ordinated, through a Science and Technology Advisory Committee, by the Scientific Secretary,

The administration of the EFC is handled by three Secretariats: DECHEMA e. V. in Germany, the Societe de Chimie Industrielle in France, and The Institute of Materials in the United Kingdom. These three Secretariats meet at the Board of Administrators of the EFC, There is an annual General Assembly at which delegates from all member societies meet to determine and approve EFC policy. News of EFC activities, forthcoming conferences, courses etc. is published in a range of accredited corrosion and certain other journals throughout Europe. More detailed descriptions of activities are given in a Newsletter prepared by the Scientific Secretary.

The output of the EFC takes various forms. Papers on particular topics, for example, reviews or results of experimental work, may be published in scientific and technical journals in one or more countries in Europe. Conference proceedings are often published by the organisation responsible for the conference.

In 1987 the, then, Institute of Metals was appointed as the official EFC publisher. Although the arrangement is non-exclusive and other routes for publication are still available, it is expected that the Working Parties of the EFC will use The Institute of Materials for publication of reports, proceedings etc. wherever possible.

The name of The Institute of Metals was changed to The Institute of Materials with effect from 1 January 1992.

A. D. MercerEFC Series Editor,The Institute of Materials, London, UK

viii Series Introduction

EFC Secretariats are located at:

Dr B A RickinsonEuropean Federation of Corrosion, The Institute of Materials, 1 Carlton House Terrace, London, SW1Y 5DB, UK

Mr P BergeFederation Europeene de la Corrosion, Societe de Chimie Industrielle, 28 rue Saint- Dominique, F-75007 Paris, FRANCE

Professor Dr G KreysaEuropaische Foderation Korrosion, DECHEMA e. V., Theodor-Heuss-Allee 25, D- 60486, Frankfurt, GERMANY

Corrosior transports associatec then depe: The prese: only give enhanced fluids and associated particular! fact C 0 2 c< encounter* applicatioi systematic design oil i

This do corrosion, results and chemistry < affect the p is on the u< materials i economy, a

This do< design engi developmei physical co corrosion cc of oilfield < transporta ti engineers, p

*'Sour corrosio

Preface

Corrosion is a natural potential hazard associated with oil and gas production and transportation facilities. This results from the fact that an aqueous phase is normally associated with the oil and/or gas. The inherent corrosivity of this aqueous phase is then dependent on the concentration of dissolved acidic gases and the water chemistry. The presence of H-5, C 0 2, brine and/or condensed water with the hydrocarbon not only give rise to corrosion, but also can lead to environmental fracture assisted by enhanced uptake of hydrogen atoms into the steel. C 0 2 is usually present in produced fluids and, although it does not cause the catastrophic failure mode of cracking associated with H0S*y its presence can nevertheless result in very high corrosion rates particularly where the mode of attack on carbon and low alloy steels is localised. In fact C O 2 corrosion, or 'sweet corrosion', is by far the most prevalent form of attack encountered in oil and gas production and is a major source of concern in the application of carbon and low alloy steels. Hence, the need to have a document which systematically addresses the steps, considerations and parameters necessary to design oil and gas facilities with respect to C 0 2 corrosion.

This document sets the scene on design considerations specifically related to C 0 2 corrosion. It has been developed from feedback of operating experience, research results and operators' in-house studies. Particular attention has been given to the chemistry of the produced fluid, the fluid dynamics and physical variables which affect the performance of steels exposed to C 0 2~containing environments. The focus is on the use of carbon and low alloy steels as these are the principal construction materials used for the majority of facilities in oil and gas production offering economy, availability and strength.

This document is a practical, industry oriented guide on the subject for use by design engineers, operators and manufacturers. It incorporates much of the recent developments in the understanding of the ways in which detailed environmental and physical conditions affect the risk of C 0 2 corrosion. It also describes means of corrosion control. It is comprehensive in addressing C 0 2 corrosion of all major items of oilfield equipment and facilities incorporating, production, processing and transportation. As such, it provides a key reference for materials and corrosion engineers, product suppliers and manufacturers working in the oil and gas industry.

"So u r corrosion', resulting from the presence of H; S, is the subject of EFC Publications Numbers 16

Acknowledgements

The CO., Corrosion Work Group of the EFC Working Party on Corrosion in Oil and Gas Production held its first meeting in September 1993. Since then, several meetings have been held to address industry-wide issues related to engineering design for CO, corrosion. The organisation of the Work Group was undertaken by representatives from worldwide oil and gas producers, manufacturers, service companies and research institutions.

In achieving the primary objective, parameters affecting C O , corrosion, its mechanism and methods of control have been discussed during the Work Group meetings. These aspects form the core of the present document, Sections of which have been prepared by the Work Group members.

The chairmen of the W orking Party and Work Group would like to thank all who have contributed their time and effort to ensure the successful completion of this document. In particular we wish to acknowledge a significant input from these individuals and their respective companies:

J Pattinson, A McMahon and D Harrop, BP, UKJ-L Crolet, Elf, FranceA Dugstadt, IFE, NorwayG Schmitt, MF1, GermanyY Gunaltun, Total, FranceE Wade, previously with Marathon, UKO Strandmyr, Statoil, NorwayW Lang, Bechtel, UKJ Palmer, CAPCIS, UKM Swidzinski, Phillips, UKM Celant, MaC, ItalyP O Gartland, CorrOcean, NorwayR S Treseder, CorrUPdate, USAT Kolt, Conoco, USAN Farmilo, AEA Technology, UK

In addition, valuable comments from R Connell and B Pots (Shell, The N eiheriands) and T Gooch (TW1, UK) are appreciated.

Finally, one of the editors (MBK) wishes to thank BP for their support and permission to publish some of the information in this document.

Bijan Kermani Liane SmithChairman o f CO-, Corrosion Chairman oj EFC Working Party onGroup Workshop Corrosion in Oil and Gas Production

C 0 2 c transp and d< there r which adequ in dee; often 1

The the lif partici movec fields i (somei export such s inhibit Never corrosi produt is cons

In t there i

Adeis

Ai th< all

Areibe

A . th< co

Introduction

C O , corrosion has been a recognised problem in oil and gas production and transportation facilities for many years. Despite systematic attempts to analyse it and develop predictive models, it is still not a fully understood phenomenon and there remains ambiguity and argument on the engineering implications of parameters which affect it. Furthermore, most of the present predictive models are not based on adequate information to take into account the increasingly harsh environments seen in deep wells and they also take little account of hydrodynamic parameters, and so often lead to conservative designs.

The problem cannot be said to be a diminishing one, since reliable prediction of the life of carbon steel components in production systems remains unclear HL particularly, in the current situation where oil and gas exploration activities have moved to more marginal areas and harsher operational conditions. Many of these fields necessitate the transportation of raw wellhead gas and fluids either from wells (sometimes subsea) or from remote areas to a central processing facility, with the export of treated fluids to a distant terminal/additional processing facility. Although such systems have often been designed to operate successfully with corrosion inhibition, there have been instances where this approach has failed in practice. Nevertheless, with detailed evaluation of the corrosion risk, combined with a proper corrosion management programme (control, monitoring, inspection and assessment), production and transportation of wet hydrocarbon gas and oil in carbon steel facilities is considered technically viable.

In brief, where there is a risk of infernal corrosion in wet production facilities there is a need for:

A design m ethodology for reviewing the potential corrosion risks and developing a suitable design and corrosion allowance where appropriate. This is the principal subject of this document.

An inhibitor deployment programme including why inhibitors are used, how they are selected and how to achieve maximum performance in the field to alleviate internal corrosion of facilities.

A corrosion control management programme w'hich, based on the design review, details the procedures for corrosion control, how such corrosion is to be monitored and how the facilities are to be inspected

A defect assessment methodology vvhich determines whether the integrity of the facility is compromised or likely to be compromised, in the event that a corrosion defect is detected.

2 C 02 Corrosion Control in Oil and Gas Production Design Considerations

In this docum ent the emphasis has been placed primarily on the first point and the other three points have been addressed briefly.

The first step in establishing the design methodology is an understanding of C 0 2 corrosion. This requires a multi-disciplinary approach, involving knowledge of fluid chemistry, hydrodynamics, metallurgy and inhibitor performance and partitioning. Mechanistic understanding of the phenomenon is essential to enable development of engineering criteria for accurate prediction of the form and rate of corrosion which may occur. This document aims to address these issues.

This doc transpo] commer the purp oil and/

The r damage forms o corrosio

T h ek 5. An un corrosio and hov wrhich d and the

Sectic of mino addition

In c o j and app; design, t with in i

The d within t Section c

Well

Proc

Gas

Final!different

ng of CO-, ge of fluid rtitioning. elopment ion which

Scope

This document sets out a proposed design philosophy for the production and pipeline transportation of wet oil, wet gas and multiphase fluids, for use in the technical/ commercial assessment of new field developments and in prospect evaluations. For the purpose of this document, wet oil, wet gas and multiphase fluids are defined as oil and /or gas containing water and C 0 2

The mechanism of CO, corrosion is explained and the forms that the corrosion damage can take are described in Section 3. This is followed by a description of the forms of CO-, corrosion damage and the steps necessary to minimise localised corrosion of carbon steel welds (Section 4).

The key parameters influencing the rate of C 0 2 corrosion are discussed in Section5. An understanding of the role of the carbonate scale in influencing the form of the corrosion is shown to be important in understanding how some inhibitors operate and how the nature of the scale changes with temperature. This leads to Section 6 which describes a summary of the models available for predicting the corrosion rate and the parameters they incorporate.

Section 7 deals with various methods of corrosion control, including the addition of minor alloying elements and changing the corrosive environment through the addition of pH controller, glycols or corrosion inhibitors.

In considering the application of this knowledge on forms of corrosion damage and approaches to corrosion rate prediction and mitigation to the question of facilities design, the first issue is to establish an appropriate corrosion allowance. This is dealt with in Section 8.

The document then highlights parameters which are significant to different items within the production facilities. For the purposes of discussing corrosion design, Section 9 has been divided into:

Well Completions;

Production Facilities (including flowlines and pipelines); and

Gas Reinjection Systems.

Finally, some comments are given on corrosion monitoring appropriate to the different facilities.

The Mechanism of C 0 2 Corrosion

The problem of CO, corrosion has long been recognised and has prompted extensive studies. Dry CO-* gas is not itself corrosive at the temperatures encountered within oil and gas production systems, but is so when dissolved in an aqueous phase through which it can promote an electrochemical reaction between steel and the contacting aqueous phase. C 0 2 is extremely soluble in water and brines but it should also be remembered that it has even greater solubility in hydrocarbons potentially 3:1 in favour of the hydrocarbon. Hydrocarbon fluids are generally produced in association with an aqueous phase. In many cases the hydrocarbon reservoir will also contain a significant proportion of COr As a result of this, CO-, will dissolve in the aqueous phase associated with hydrocarbon production. This aqueous phase will corrode carbon steel.

Various mechanisms have been postulated for the corrosion process but all involve either carbonic acid or the bicarbonate ion formed on dissolution of C 0 2 in water this leads to rates of corrosion greater than those expected from corrosion in strong acids at the same pH. CO, dissolves in water to give carbonic acid, a weak acid compared to mineral acids as it does not fully dissociate.The steps of carbonic acid reaction may be outlined as follows:

C 0 2(1 + H p ^ C 0 2ldls ,v>lvedl + H2 H 2 C O ? + H C O r (2)

The mechanism postulated by de Waard [2-4] is, perhaps, the best known:

H2C 0 3 + e ^ H + HCO,~ (3)

2 H 4 H~, (4)

with the steel reacting:

Fe Fe2_r + 2e~ (5)

and overall:

CO, -r HX> -f Fe > FeCO- (iron carbonate) + H, (6)

Whilst there is some debate about the mechanism of CO, corrosion in terms ofwhich dissolved species are involved in the corrosion reaction, it is evident that the

The Mechanism o f C 0 2 Corrosion 5

resulting corrosion rate is dependent on the partial pressure of C 0 2 gas. This will determine the solution pH and the concentration of dissolved species.

In reality, the complete chain of electrochemical reactions is much more complex than this brief outline. Depending upon which is the rate determining step the dependance of the electrochemical reactions on pH and dissolved C 0 2 varies.

I extensive red within se through contacting ild also be ially 3:1 in issociation o f 'tain a tev^ nueous ill corrode

all involve in water n in strong weak acid

(1)

(2)

/n:

(3)

(4)

(5)

(6)

in terms of ent that the

Types of C 0 2 Corrosion Damage can Fedge:betw

FI.C 0 2 corrosion may manifest itself as general thinning or localised attack. Localised orcorrosion is characterised by loss of metal at discrete areas of the surface withsurrounding areas remaining essentially unaffected or subject to general corrosion. distuThese discrete areas may take various geometrical shapes. Thus, circular depressions m a y !usually with tapered and smooth sides are described as pits. Stepped depressions protewith a flat bottom and vertical sides are referred to as mesa attack. Other geometricalforms of localised corrosion include slits (sometimes referred to as knife line), groovesetc. In flowing conditions localised attack may take the form of parallel groovesextending in the flow direction; this phenomenon is known as flow induced localisedcorrosion. Local

experparth

4.1. Localised Corrosion of Carbon Steel the mgeom

CO, corrosion can appear in three principal forms, pitting, mesa attack or flow Iniinduced localised corrosion. to cor

Pitting can occur over the full range of operating temperatures under stagnant to the p?moderate flow conditions. The susceptibility to pitting increases and time for pitting Thonic acid arising iect reaction with iirectly related to us phase. Thus in pressure of C 0 2 in

0 2 content of the ictw ith the fluids t t bw pressure

us phase that will >erse. This activity on of the aqueous ectly linked to the t gas is effectively al pressure, al gas will play an ;acity /COi should

(7)

ive estimate for /. gacity coefficient, tin any predictive

)th ^as and water lit the aqueous his is particularly

nment can lead to ;gests that an iron steel, a carbonate 171. The formation

. that are described

Total system pressure, bar

Fig. 1 Fugaciiy coefficient for CO, in methane for gas mix fares with less than 5 moie% C 0214}.

H ow ever, at h igher tem peratures (e.g . around 80C ) the iron carbonate solubility is decreased to such an extent that scale form ation is more likely. Under laboratory conditions, rates of uniform corrosion are consistently reduced at higher tem peratures.

Some laboratory studies show that the initial rate of uniform corrosion increases up to 70-90-C, probably due to the increase of mass transfer and charge transfer rates [2,33. Above these temperatures, the corrosion rate starts to decrease. This is attributed to the formation of a more protective scale due to a decrease in the iron carbonate solubility and also to the competition between the mass transfer and corrosion rates. As a result, a diffusion process becomes the rate determining step for the corrosion rate.

Field evidence for a m axim um tem perature for C 0 2 corrosion has been found in some wells. These case histories show that in oil and gas wells maximum corrosion takes place where the tem perature is betw een about 60 and lOO^ -'C [2,18,191 which may coincide with dew point tem perature in gas wells. In these cases, below 60-70-C , the corrosion rate increased with increasing tem perature and above 8 0 -1 0 0 ~C the corrosion rate decreased with increasing temperature. Conversely, very high corrosion rates have been observed up to 130C at the top of som e gas wells exascerbated by high rates of w ater condensation.

14 C 07 Corrosion Control in Oil and Gas Production Design Considerations

5.4. pH

The pH value is an important parameter in corrosion of carbon and low alloy steels. The pH affects both the electrochemical reactions and the precipitation of corrosion products and other scales. Under certain production conditions the associated aqueous phase can contain salts which will buffer the pH. This tends to decrease the corrosion rate and lead to conditions under which the precipitation of a protective film or scale is more likely.

For bare m etal surfaces which are representative for worst case corrosion, laboratory experiments indicate that a flow sensitive H+ reduction dominates the cathodic reaction at low pH (pH < 4.5) while the amount of dissolved CO., controls the cathodic reaction rate at higher pH (pH > 5).

In addition to the effects on the cathodic and the anodic reaction rates, pH has a dominant effect on the formation of corrosion films due to its effect on the solubility of ferrous carbonate, as illustrated in Fig. 2. It is seen that the solubility of corrosion products released during the corrosion process is reduced by just five times when the pH is increased from 4 to 5 but by a hundred times with an increase from 5 to 6. The lower solubility gives a much higher F eC 03 supersaturation on the steel surface and a subsequent acceleration in precipitation and deposition of iron carbonate scale [17]. The likelihood of protective film formation is therefore increased significantly when the pH is increased beyond 5 and this can explain why low corrosion rates have been reported for many fields where the pH is in the range 5.5-6. However, the solubility of F e C 0 3 must not be confused with that of ferrous ions (Fe2").

Reliance < protectior scale brea

Recent the initial Corrosion resulting i of such a si (which wi underlying

The sea of organic smooth flc this; only occurrence from the s coincident With the su arises at sui scale prote<

The effe laboratory t care espt

pH

Kg.2 Solubility of iron carbonate released during the corrosion process at 2 bar C 02 partial pressure and 40 C [17].

Leaving asi< H2S can hav can give prc mean that fa a lower rate partial press

The acid i than carboni result, the co: similar. H2S ; rate which is in the type * decreasing tl

Many pap under ambie and gas indu H2S and C 0 2

ons Key Parameters Affecting Corrosion 15

v alloy steels, i of corrosion le associated > decrease the f a protective

se corrosion, ominates the C 0 2 controls

tes, pH has a tb( 'liability / of corrosion ? times when e from 5 to 6. i steel surface rbonate scale significantly

>rrosion rates However, the -+).

6

7

~0, part h i

5.5= Carbonate Scale

Reliance on carbonate scales/film as described in section 5.3 to give continuous protection is not totally warranted. In particular, in regions of high flow or at welds, scale breakdown can lead to rapid rates of localised corrosion ('m esa attack').

Recent extensive work on the subject has shown that the corrosion process involves the initial production of an iron carbide matrix on the surface of corroding steel. Corrosion product film of F eC 03 or Fe30 4 will then form as a scale on the surface resulting in a reduction in the corrosion rate [201. The formation and protectiveness of such a scale depends on a number of factors such as the solubility of iron carbonate (which will vary with pH and the presence of other salts), the rate of reaction of the underlying steel and the surface condition (roughness/cleanliness/prior corrosion).

The scale [91 may be weakened by high chloride concentrations, by the presence of organic acids or it can be eroded by high speed liquids. Practical velocities for smooth flow in systems with single phase liquid flow are often too low to achieve this; only the im pact of high speed liquid droplets can dam age the scale. The occurrence of such a disturbed flow pattern in practical systems can be predicted from the suggestion made by Sm art [211 that the onset of erosion-corrosion is coincident with the transition to the annular mist flow regime in multiphase flow. With the superficial liquid velocities associated with wet gas transport, this transition arises at superficial gas velocities between 15 and 20 ms1. Above these velocities the scale protectiveness may be impaired.

The effects of short term scaling will often make interpretation of short-term laboratory experiments difficult and for this reason such data must be treated with care especially results that give unexpectedly low rates of corrosion.

5.6. The Effect of H2S

Leaving aside the cracking and corrosion problems associated with sour service, H2S can have a beneficial effect on wet hydrocarbon C 0 2 corrosion as sulfide scales can give protection to the underlying steel. The effect is not quantified but it does mean that facilities exposed to gas containing low levels of H2S may often corrode at a lower rate than completely sweet systems in which the temperatures and C 0 2 partial pressures are similar.

The acid formed by the dissolution of hydrogen sulfide is about 3 times weaker than carbonic acid but H25 gas is about 3 times more soluble than C 0 2 gas. As a result, the contributions of CO, and H2S partial pressures to pH lowering are basically similar. H2$ may cause corrosion also in neutral solutions, with a uniform corrosion rate which is generally very low [221. Furthermore, H^S may play an important role in the type and mechanical resistance of corrosion product films, increasing or decreasing their strength.

Many papers have been published on the interaction of H2S with low carbon steels under ambient conditions and the work relating to H2S corrosion problems in the oil ^nd gas industry is w'ell documented. However, literature data on the interaction ofH,S and C 0 2 is still limited. The nature of the interaction of K ,S and C O , with carbon

16 CO-, Corrosion Control in Oil and Gas Production Design Considerations

steel is complex. From past experience corrosion product layers formed on mild steel can be protective or can lead to rapid failure depending on the production conditions. This is primarily because an iron sulfide (FeS) film will form if HUS is predominant and iron carbonate (F eC 03) will form if CO., is predominant in the gas.

The majority of the open literature does indicate that the CO., corrosion rate is reduced in the presence of H ,S at am bient temperatures. However, it must be emphasised that H^S may also form non-protective layers [23], and that it catalyses the anodic dissolution of bare steel 1241. There is a concern that steels may experience some form of localised corrosion, but very little information is available.

Published laboratory work has not been conclusive, indicating that there is a need to carry out further study in order to clarify the mechanism [25,26]. A recent failure showed how the corrosion rate in the presence of a high concentration of H2S may be higher than predicted using CO0 corrosion prediction models [271. However, in spite of the work on H2S corrosion of steels, no equations or models are available to predict corrosion as is the case for C O , corrosion of steels.

Cracking of metals in production environments containing H2S is a major risk. Hydrogen sufide can cause cracking of carbon and low alloy steels within certain conditions of H2S partial pressure, pH, temperature, stress level and steel metallurgy and mechanical properties (e.g. hardness). The type of damage manifests itself in the form of cracking such as sufide stress cracking (SSC), stepwise cracking and other forms of damage which are discussed at greater length in EFC Publication No.16.

5.7. Wax Effect

The presence of wax in main oil lines can influence C 0 2 corrosion damage in two ways; exacerbating the damage or retarding it, the effects depending on other operational parameters such as temperature, flow, etc. and uniformity and the nature

-of the w ax layer.Field experience in sweet oil lines in the USA, have shown that a layer of wax

(paraffin) deposited on a carbon steel surface can result in severe pitting of the steel in anaerobic aqueous solutions of carbon dioxide 128]. Severe pitting occurred along the bottom of the pipe. Pitting (small random pits) tended to concentrate at the start of an uphill run where water could collect. Scale analysis showed the presence of iron sulfide. This was attributed to the presence of bacteria. (The detection of sulfide in a swreet oil line is not usual. In fact in the case of microbially assisted corrosion, scale analyses often show 15-30% FexSv). Velocity was an apparent factor affecting ; the location of pits; there being a decrease in the number of pits at flow velocities above about 0.6 m s '1. (The principal practical observation was that conventional commercial corrosion inhibitors were ineffective in controlling corrosion; the corrosion control measure finally adopted for the gathering lines was to install pull-through polyvinyl chloride liners). In this case the proposed corrosion mechanism is of diffusion of carbon dioxide through the wax layer which is thought to provide a large cathodic area that supports anodic dissolution of the steel at discontinuities of the wax layer. The effect was reproduced in laboratory tests with paraffin coated specimens exposed to C 0 2 saturated water at atmospheric pressure and ambient

tempera t The area: difficulty inhibitor:

In con' length) si attributec corrosion injected i: confirmee

Key Parameters Affecting Corrosion 17

i mild steel :onditions. edominant

don rate is it must be it catalyses experience

re is a need :ent failure >f H2S may fowever, in lvailable to

temperature. Localised corrosion only took place where there was no wax deposit. The areas covered with wax were protected from the C 0 2 containing solution. The difficulty in controlling this type of localised corrosion with commercial oilfield inhibitors was demonstrated in these laboratory tests [281.

In contrast field experience of a 20 in. (50.8 cm) oil line in Indonesia (about 20 km length) showed almost nil corrosion rate during about 10 years service which was attributed to a wax deposit on the pipe wall. The water cut was up to 50%. Internal corrosion started when light hydrocarbon condensate produced from a gas field was infected into the line. This dissolved the wax deposit exposing the steel surface, as confirmed by internal inspection of a corroded pipe section.

major risk. ;hin certain metallurgy sts itself in acking and lication No.

lage in two ig on other d the nature

lyer of wax ; of the steel urred along ea{ '-? start presence of >n of sulfide d corrosion, tor affecting w velocities onventional he corrosion Hill-through tanism is ofo provide a ntinuities of affin coated nd ambient

Prediction of the Severity of C 0 2 Corrosion

It is apparent that CO-, corrosion of carbon and low alloy steels has been, and remains, a major cause of corrosion damage in oil and gas field operations 11], The industry relies heavily on the extensive use of these materials, and thus there is a desire to predict the corrosivity of C 0 2-containing brines when designing production equipment and transportation facilities.

A true industry standard approach to predicting C 0 2 corrosion does not exist although there are aspects of commonality between the approaches /models offered by a number of operators, research organisations and academic establishments. Apart from limited reference in National Gasoline Association of America [29] and American Petroleum Institute [30] publications, there is no professional body or agency to provide a standard guideline on C 0 2 corrosion prediction. However, in particular, the work of Shell in this area has provided a reference point. The Shell (de Waard et al.) equation or nomogram has been developed as an engineering tool. It presents, in a simple form, the relationship between potential corrosivity (worst case) of aqueous media for a given level of dissolved CO defined by its partial pressure, at any given temperature. The relative simplicity of the Shell approach and its ease of use have undoubtedly been positive factors in its broad acceptance. This is in contrast to the arguably more 'all-encompassing' models of, for example, Southwestern Louisiana, VHRITEC, CAPCIS and others which require more detailed input data to run them. Also input of inspection/monitoring data may be called for to refine the models' accuracy or field/well specificity.

There would appear to be a trade-off between a model's relative ease of use versus availability, detail and reliability/accuracy of necessary input data/conditions combined with the degree of accuracy/absoluteness required in the assessment of the corrosion risk. The last will also be influenced by the ease and sensitivity of subsequent corrosion monitoring and inspection.

There still remains an absence of any strong systematic correlation betw een predicted and actual field corrosion rates and experience, although CORMED goes someway in this respect [311. Future development of predictive models should contain a much stronger element of field correlation.

The engineer ideally wants a predictive tool that can be readily applied and is suitable for application at all stages of project development and subsequent operation. This may seem a tall order but it may nevertheless be argued that the fundamentals of the C 0 2 corrosion process will be common to all situations; It is the overlying effects of such factors as flow regime, film formation/deposition, hydrocarbon phase and corrosion inhibitor which cloud or complicate the picture. Both the Shell and CORM ED models have been developed from a basic consideration of the CO., corrosion reactions, the former more empirical in origin and the latter more theoretical. Both have then attempted to account for the overlying effects either by applying correction factors (Shell) or through field correlation (CORMED).

not to } decisio: corrosi* para me thereto;

Base damagt betweei conditic what is subseqi inspecti latter, u

Figui define a flow reg to achie

Differenpredictk

wn w tw een *M ED goes mid contain

>lied and is it operation, nda mentals e overlying irbon phase ie Shell and of the C 0 2 .* theoretical. >y applying

Notwithstanding the above discussion, the intent of the present document was not to provide or recommend a particular corrosion prediction tool, but leave the decision to the individuals. Nevertheless, this section provides an overview of CO, corrosion models and parameters considered in each model. Furthermore, the parameters which are considered essential in designing for C O , corrosion and are therefore needed, no matter which predictive tool is used, are presented in Fig. 3.

Based on the foregoing discussion, the procedure for predicting C 0 2 corrosion damage is described in Fig. 4. A key feature is the positive and ongoing interaction between the corrosion engineer and petroleum engineer to ensure that relevant service conditions are defined and detailed. There has to be a common understanding of w hat is required against the lim itations of the selected predictive model and subsequent monitoring/inspection. A case is made for rationalising monitoring and inspection data with predicted rates, to strengthen the relevance and validity of the latter, whilst working to introduce a stronger predictive element to the former.

Figure 5 summarises the necessary overall critical steps identified in working to define a risk of C 0 2 corrosion. It should also be recognised that characterising the flow regime/shear stress to establish water wetting (Section 5.1) may also be criticar to achieving effective corrosion inhibitor selection and deployment (Section 7.4).

6.1. C 0 2 Corrosion Prediction Models For Carbon Steel

Different oil companies and research institutions have developed a large number of prediction models. Table 1 (p.22) gives an overview of the parameters treated in

Hydrodynamics:Local/bulk flow regimes

Top of line/Bottom of line

Acid gases:co2(H2S)

Steel:Composition

Microstructure Weid; composition, profile

Fluid chemistry:Local/bulk analyses pH, organic acids

Controlling ParametersMicro-alloying elements

Corrosion inhibition GJycoi and methanol

pH-control

Operating condition:Temperature, pressure

Number of phases, water cut (over the life of the field)

Others:Initial production condition

Trend of water cut Carbonate scale Scale inhibitor Other additives

Fig. 3 Parameters affecting CO, corrosion design.

20 CO-, Corrosion Control in Oil and Gas Production Design ConsiderationsCOMMENTS

Specific case

PETROLEUMENGINEER

Water analysis Total P or Bubble Point Temperature mole% C02 H2S present?

Flow Regime Analysis

CORROSIONENGINEER

SERVICECONDITIONS

CONSIDERCHEMISTRY

EFFECT

RATIONALISE1 {vs monitoring ,___ j1 and/or inspection | data) i

CORROSIONDAMAGE/RATE

Positive interaction at ali times.

Consider total life ol the field.

Check on solution pH. Validate measured pH.

Worst case corrosion rate. Erosion not considered. {Oil/water ratio/flow regime need to be considered, cf. water or oil wetting.)

Check sensitivity to velocity.

Does not predict corrosion rate in presence of H2S.

Determine total accumulative corrosion damage over field life.

Fig. 4 Procedure for predicting C 02 corrosion damage for a given water composition, CO-, partial pressure and temperature.

those models which have been fully or partly described in the literature. It is seen that different parameters are used as inputs and it is also seen that some of the key parameters listed in Fig. 3 are not included at all.

Very different results are obtained when the models are run for the same test cases. This is due to the various philosophies used in the development of the models. Some of the models give a worst case corrosion rate based on fully water wetting and little protection from scale and inhibitors. These m odels have a built-in conservatism and they probably over-predict the corrosion attack significantly for many cases. Other models are partly based on field data and predict generally much

Fig. 5 C

lower con or form al to less tha

The me ah). The fi [2]. The m of pH anc model wa taken into in 1995 [3 generated

Prediction of the Severity of CO, Corrosion 21

mes.

eld.

ne

city,

on rate

dative field life.

O, partial

e. It is seen e of the key

e same test the models, iter wetting e a built-in ificantly for erally much

[ monitoring/inspection j

CORROSION DAMAGE/RATE

Fig. 5 Critical steps in defining CO, corrosion damage.

Sower corrosion rates. In these models it is assumed that reduced water wetting and/ or formation of protective scale can reduce the corrosion rate from many mm/year to less than 0 .1.

The most frequently referenced model has been developed by Shell (de Waard et itl.). The first version, based on temperature and P co, only, was published in 1975 (21. The model has since been revised several times. Correction factors for the effect of pH and scale were included in 1991 [32j. To account for the effect of flow a new model was proposed in 1993 where the effect of mass transport and fluid velocity is taken into account [3]. A revised version including steel composition was published in 1995 [331. This model represents a best fit to a large number of flow loop data generated at IFE [341.

Table I. An overview of the parameters treated in the various prediction models

Models

Parameters Shell 75 Shell 91 Shell 93 Shell 95 CORMED UPUCOR SSH ksc m) USL PREDICT

f\:o2 t

Temperature t f

pH t

Flow rate

Flow regime 0 t

Scale factor t 0

pM 0

Steel [x!

Water wetting M g) [xj E) t

Ca/ HCO,

H,S

HAc

Field data t

Ref. 2 32 3 33 31 35 36 37 38 ?g

9 Parameters considered directly[xj Parameter considered indirectly or not considered highly influential.

Prediction of the Severity of C 02 Corrosion 23

The CORM ED model developed by Elf predicts the probability of corrosion in wells [31]. It is based on a detailed analysis of field experience on C 0 2 corrosion mainlv from Elf's operations, but also from data supplied or published by others (e.g. Total, Phillips). The model identified the C 0 2 partial pressure, in situ pH, Ca2V HCCX- ratio and the amount of free acetic acid as the only influencing factors for downhole corrosion and predicts either a low risk, medium risk or a high risk for tubing perforation within 10 years.

The LIPUCOR corrosion prediction program calculates corrosion rates based on tem perature, C 0 2 concentration, w ater chemistry, flow regim e, flow velocity, characteristics of the produced fluid, and material composition [35]. The program which is developed by Total is based on both laboratory results and field data. More than 90 case histories have been used in the development.

The SSH model is a worst case based model mainly derived from laboratory data at low temperature and a combination of laboratory and field data at temperatures above 100C [36]. The model has been developed by Hydro, Saga and Statoil in collaboration with IFE.

IFE is developing a new predictive model for CO , corrosion based on mechanistic * m odelling of electrochemical reactions, transport processes and film formation processes. The first part of the model which applies for the case when no surface films are present has been published recently [37].

The USL model predicts corrosion rates, temperatures, flow rates, etc. for gas condensate wells [381. It is a package of programs developed by University of Southwestern Louisiana.

Predict is a software tool developed by CLI international [39], The basis of the model the de Waard-Milliams relationship for C 0 2 corrosion, but other correction factors are used and a so-called 'effective C 0 2 partial pressure' calculated from the system pH.

C 0 2 Corrosion Control

CO-> corrosion damage and its severity can be mitigated by a number of measures. These primarily fall into two broad categories of (i) modifications to carbon and low alloy steels, to enhance their resistance to corrosion, and (ii) alteration of the environment to render it less corrosive.

7.1. Micro-alloying of Carbon and Low Alloy Steels

Much work has been done to try to improve the corrosion resistance of carbon and low alloy steels with small additions of alloying elements. The corrosion rate is controlled by the transport of the reacting agents through the corrosion product layer and the different alloy additions may affect the protectiveness of the surface film. The microstructure of the steel is also important. It is apparent that the alloying elements and the microstmchire do not necessarily have the same effect when the steel is exposed at a low pH, in formation w*ater, in injection water or in inhibited solutions or wThen different corrosion products accumulate at the steel surface. This may be the reason why there is conflicting information on this subject in the literature.

Note that the control of corrosion in carbon steel welds was discussed in Section4.2.

7.1.1. Effect of Chromium

Chromium is the most commonly used alloying element added to steel to improve the corrosion resistance in wet C 0 2 environments. Independent work at Sumitomo [40], Kawasaki [41] and IFE [42] shows a beneficial effect of small amounts of chromium in C 0 2 saturated water at temperatures below 90C. It is suggested that Cr is enriched in the iron carbonate film and makes it more stable. Alloys with 0.5% Cr seems to be a good choice giving good corrosion properties and hardly any loss of toughness.

At higher temperatures the effect of chromium seems to be more unclear and several authors have reported a reduction in corrosion resistance above 100C for low alloyed chromium steels [5,43,44]. In contrast it has also been reported that the temperature giving a maximum corrosion rate increases with increasing Cr content in the steel [40].

Field experience does indicate an improvement of the corrosion resistance with small amounts of chromium and several companies have recently specified 0.5-1% Cr for their pipelines.

Thepartcan

1

Be focus ferrit phas' (low agah and f a goo fine ft at a h

Qvcarbo for th. the cl m icro forma carbid on the

Sint will hi

Nickel improA disagn M ost r< also be

A p t reporte

C0^ Corrosion Control 25

easures. and low i of the

bon and n rate is uct layer ace film, alloying vhen the nhibited ace. This iterature. a Section

improve air^om o lOv.. & of 'Sted that .vith 0.5% /any loss

clear and 100C for d that the 'r content

ance with _>d 0.5-1 %

7.1.2. Effect of Carbon

The effect of carbon is linked to the carbide phase, cementite (Fe3C) which forms part of the microstructure of carbon steels. There are two effects of cementite that can be emphasised:

Iron carbide is exposed at the steel surface when the iron is dissolved and it then causes an increase in the corrosion rate. This is explained by a galvanic effect where the cementite acts as a cathode.

The cementite can act as a framework for build-up of a protective corrosion film.

Both these points are connected to the microstructure. The literature is mainly focused on ferrite-pearlite structures and quenched and tempered (QT) steels. A ferrite-pearlite structure can form a continuous grid of cementite after the ferrite ^phase is removed by corrosion. Under conditions where film formation is impeded (low temperature and low pH) this carbide phase increases the corrosion rate due to a galvanic coupling between the cementite and the ferrite leading to local acidification and further difficulty in establishing protection. Such a grid of carbide could also be a good anchor for a protective iron carbonate film under film forming conditions. A fine ferrite-pearlite structure will improve this tendency. These effects will be stronger at a high carbon content (> 0.15% C).

Quenched and tempered steels contain mainly martensite or bainite where more carbon is in solid solution and the carbide phase does not make a continuous grid as for the ferritic-pearlitic steels. In these steels the galvanic effect will be reduced and the chance of anchoring a protective film less. M ost reports on the effect of microstructure m aintain that ferrite-pearlite is favourable with respect to film formation [43,45-471 while other workers suggest that QT steels with needle-like carbides can anchor a film better than a ferrite-pearlite steel [44]. This might depend on the very first period of exposure.

Since new pipeline steels have low carbon content (< 0.1 % C); the effect of cementite will be of less importance in these types of steels.

7.1.3. Effect of Other Alloying Elements

Nickel is often added to the steels and in welding electrodes for pipeline steels to improve weldability and the toughness of the weld deposit. There has been some disagreement about the effect of small amounts of nickel on COz corrosion [41,42,48]. Most reports indicate a negative effect, but it seems to be slight Varying effects have also been reported in different sources with small additions of copper [41,44,48].

A positive effect of molybdenum [49], silicon [44,49] and cobalt [39,49] has been reported, but a more systematic study is required to confirm this.

26 CO, Corrosion Lontro! in Oil and Gas Production Design Considerations

7.2. Effect of Glycol and Methanol

Large quantities of glycol or methanol are often introduced into wet gas-producing system s to prevent and control hydrate form ation which can cause plugging problems. Both of these chemicals, if present in sufficient concentrations can inhibit C 0 2 corrosion. Of the two, glycol is much more effective and a correction can be made to the predicted corrosion rate to take this into account. Combined with a pH controlling agent, the water/alcohol phase may be rendered less corrosive (Section 7.3).

The glycol additives which are mainly used for hydrate prevention are MEG (mono-ethylene glycol) and DEG (di-ethylene glycol), but TEG (tri-ethylene glycol) can also be used for dehydration. These are effective in reducing the rate of CO., corrosion by diluting free water and reducing the corrosivity of the resulting water phase

Methanol, too, can effectively suppress the rate of wet C 0 2 corrosion in wet gas transmission systems although it is more difficult to use in the design of corrosion protection of gas pipelines. Operators of wet gas pipelines in the UK Sector of the North Sea have found that with controlled additions of methanol carbon steel corrosion rates can be maintained below 1 mpy (0.025 mm/y) provided a methanol excess is used. For effective control the concentration of methanol in water at the pipeline reception facilities needs to be kept in excess of 80%.

Although some operators do use glycol as a means of controlling C 0 2 corrosion, this is not a recommended practice by others, as corrosion inhibition is preferred and the two effects are not norm ally considered additive (in some cases less concentrated glycol is used with inhibition). However, it is important to consider the effect that glycol carry-over from drying systems can have in an otherwise 'dry' pipeline. The glycol may absorb any residual water (further lowering the pipeline gas dewpoint) and in doing so create a water-glycol phase which could sustain corrosion, albeit at a low rate.

When evaluating corrosion protection by glycol addition, the actual composition -of the condensed glycol/water mixture is of prime importance. Models are used for these predictions, but there are no global models available which can predict all possible situations with respect to carbonate and sulfide films and the corrosion protection levels along wet hydrocarbon pipelines. The commonly used model for design with glycol effects in C 0 2 corrosive wet gas pipelines and other systems, is the Shell model [31. In normal flowing conditions the glycol/water mixture will always be in an equilibrium with the wet gas. Condensation may take place along a pipeline on the relatively colder pipewall in the top section. Nevertheless, the condensing phase will then have the same water content as the stratified glvcol, thus reducing its corrosivity.

The pH should be controlled to obtain non-corrosive conditions. In the higher pH ranges above 7-8 , the corrosion of carbon steel cannot propagate. Different pH controlling products can be used for this purpose. However, in waters containing calcium or magnesium, there is a risk for scale precipitation at higher pH values and pH control will then be impractical. Similarly, organic acids, e.g. acetic acid etc., can reduce the buffer capacity and hence the pH.

To be cost-effective and environmentally acceptable, it is standard practice to

regen< glycol pH sU: TEA.

A ci the dat can res this frc protect additic on th e ,

Theshouldprograj

As a dif universe water ch waters (i the initio 52J.

In long s1 been imp although they wou

Various clr lines are rt Historical! Further de

NaMBlong te like co

MDEA has a l(

C 02 Corrosion Control 27

producing P e g g in g :an inhibit on can be with a pH e (Section

are MEG ne glycol) ite of C 0 2 ting water

u y t gas cc*..osion

ctor of the rbon steel methanol

ater at the

corrosion, preferred cases less >nsider the wise 'dry ^\e pipeline ild sustain

imposition re used for predict all .* corrosion

rr * ?1 for systems, is ixture will ace along a heless, the ied glycol,

higher pH fferent pH containing values and :id etc., can

practice to

regenerate (i.e. reboil) the glycol/methanol after use in a system. Over time, the glycol may be partially decomposed and the pH value may decrease. In such a case, pH stabilising to obtain a system pH > 6 is necessary. Possible agents are MDEA or TEA.

A combination of glycol and corrosion inhibitors is sometimes used. As many of the data available on corrosion predictions are laboratory data, a total risk evaluation can result in the need to plan for corrosion inhibitor injection and even implement this from start-up. A question which then arises is how much additional corrosion protection the corrosion inhibitor can give. Laboratory data indicate up to 50% additional corrosion reduction, but this level of corrosion control will be dependent on the actual glycol concentration and type of inhibitor in the system.

The method of using glycol treatm ents to control C 0 2 corrosion in the field should be com bined with corrosion m onitoring and intelligent pig inspection program m e.

7.3. pH Control

7.3.1. The Role of pH

As a dissociation product of the water molecule, H+ (or its counterpart OH~) is universally involved in the kinetics of aqueous corrosion, and in the equilibria of water chemistry. The pH control or buffering by the natural alkalinity of produced waters (if any) is thus a key issue for the prediction of the C 0 2 corrosion rate (both the initial corrosion rate of bare metal, as well as the long term corrosion rate) [50 52].

7.3.2. Wet Gas Transportation Lines

In long sweet natural gas transmission lines, pH control of hydrate preventors has been implemented successfully [53]. This is a cost effective option to control corrosion, although subject to the absence of Ca2+ or Mg2+ ions in the formation water (since they would cause precipitation of scale if pH controllers are added).

7.3.3. D ifferent Chem icals and Their M echanisms

Various chemicals that have been used in operation to control the pH in natural gaslines are reviewed in this Section. Alkaline additives have changed over the years.Historically, the technique was developed by Elf in Italy (1970s) and Holland (1980s).Further developments have been as follows:

NaMBT (Sodium mercaptobenzothiazole) was used in glyco. However, in the long term it does lead to gunking problems through precipitation of a resinlike compound.

MDEA (methyldiethanolamine) was also used in glycol in the later 1980s. It has a lower freezing point than NaMBT and has no secondary effects.

28 CO, Corrosion Control in Oil and Gas Production Design Considerations

N a ,C 0 3. 1 OHXX(sodium carbonate or 'soda ash'), which may be used eitherwith glycol or methanol, is the proposed new additive as it interacts directly with the CO,/HCO-T equilibrium (50],

All pH controllers remain with the liquid phase during the regeneration of the hydrate preventer by reboiling.

The present understanding of the beneficial effect of pH control is that high pH conditions decrease the solubility limit of siderite (F eC 0 3), thus favouring the establishment of highly protective corrosion layers. Consequently, the effect of pH is nearly the same for all chemicals (NaMBT, MDEA, N aH C 03) and all solvents (MeOH, MEG, D E G ,.... or fresh water).

The in situ pH should be buffered to about 6.5, whatever the system and temperature being considered. It is worth noting that pH is here an index of the buffering level, which is the same at any temperature. Therefore, pH is measured and reported only at room temperature, whereas corrosion rates, of course, are measured at all the temperatures met along the pipeline.

73 .4 . pH Monitoring

Acetate is not a buffer for carbonic acid [54], and there is a progressive shift of the in situ pH in the presence of free acetic acid, which must be compensated by adding some fresh pH controller. Therefore, there is a need for a periodic monitoring of pH in order to detect and correct any pH shift. This is a simple pH measurement, in a sample where pure C 0 2 is bubbled under ambient condition (1 bar) in the presence of the intended chemical. This laboratory measured pH5 can be used to determine the in situ pH under pressure by:

PH fPco, > - PH i lo8 Pco, (8)

^ It is suggested to monitor this on a weekly basis for the first month after start up, and then on a monthly basis.

7.4. Corrosion Inhibition

Corrosion inhibitors continue to play a key role in controlling corrosion associated with oil and gas production and transportation. This primarily results from the industry's extensive use of carbon and low alloy steels which, for many applications, are ideal materials of construction, but generally exhibit poor C 0 2 corrosion resistance. C learly econom ics also has a m ajor part to play in m aterials selection. As a consequence, there is a strong reliance on inhibitor deployment for achieving cost effective corrosion control, especially treating long flowlines and main oil lines.

7.4.1. Inhibitor Mechanism

Corrosion inhibitors used in hydrocarbon transm ission lines are long chain com pounds. G enerally these are n itrogenous (eg. am ines, am ides, im id es,

imid; polar grou) envir* onto adsoi Howt the in much leads

Coi cathoc organ mecha m odif proces contro

The subse< enviro surfact and cc demuh a wide import in a g iv limits, recomn

Inhil on the c scale an The pre the criti

Incor also leac intensiti

For an i under t) protecti effective effective inhibitor

The i inhibito:

CO, Corrosion Control 29

eitherirectly

a of the

high pH jrin g the :t of pH is 5 (MeOH,

>tem and lef ' c the measured rnrse, are

ft of the in yy adding ing of pH ment, in a 3 presence determine

(8)

r start up,

- J

associated s from the plications, resistance, fion As a ieving cost il lines.

ong chain s, im ides,

imidazolines), but they can also be organophosphates. These compounds are either polar or ionised salts with the charge centred on the nitrogen, oxygen or phosphorus groups and as such they will be surface active. A metal surface in an aqueous environment will have a surface charge and the inhibitor will rapidly be adsorbed onto the metal surface. This process is rapid and reversible (the concentration of adsorbed inhibitor will rapidly decrease if the local environm ent is depleted). However, once adsorbed in this manner (physisorption) charge transfer between the inhibitor and the metal occurs resulting in a form of chemical bonding which is m uch more stable i.e. the inhibitor is chemisorbed. The process of chemisorption leads to the formation of a stable inhibitor film on the surface.

Corrosion is an electrochemical reaction which takes place at various anodic and cathodic sites on a metal surface the presence of an inhibitor film of long chain organic com pounds depresses both the anodic and cathodic reactions. The mechanisms are not fully clear but as well as providing a physical barrier the inhibitor modifies the surface potential and consequently limits the adsorption-desorption processes and reaction steps that occur in both anodic and cathodic reactions thus controlling corrosion.

The whole process is critically dependent on both the initial physisorption and subsequent chem isorption processes. These are strongly dependent on the environment (e.g. pH, temperature and liquid shear stresses), the state of the metal surface (e.g. roughness, scales, oxide films, surface damage and carbonate films) and com petition from other surface active species (e.g. scale inhibitors and demulsifiers). The last is particularly important in oil and multiphase systems where a wide range of oil-field chemicals may be employed. When selecting inhibitors it is important to carry out full compatibility trials to confirm that the different chemicals in a given package do not detrimentally effect each others performance beyond certain limits. Similarly, in linked systems (e.g. branch lines into a main trunk line) it is recommended that only one inhibitor be used for all of the fluids in the system.

Inhibitor molecules adsorb, however, not only on the bare metal surface but also on the carbonate scale [55]. Thus, the morphology and degree of crystallinity of the scale and, hence, its porosity (homogeneity) will be influenced by adsorbed molecules. The presence of effective inhibitors thus decreases the intrinsic stresses and increases the critical strains for cracking and spalling of the scale [56].

Incorporation of inhibitors in the surface scale and adsorption of inhibitors on it can also lead to drag reducing effects, i.e. to a reduction of wall shear stresses and local flow intensities created at flow imperfections (e.g. pits, grooves, weld beads etc.).

7.4.2. Inhibitor Efficiency and Inhibitor Performance

For an inhibitor to work effectively it must be dispersed to all wetted surfaces and under the system conditions it must be sufficiently effective to provide adequate protection. Calculations of corrosion allowances for given design lives assume effective dispersion and a certain level of success. Areas which cannot be inhibited effectively (e.g. tees) will either have to be clad or allowance made for reduced inhibitor effectiveness.

The inhibitor effectiveness can be defined in two wavs, inhibitor efficiency or inhibitor performance.

30 CO, Corrosion Control in Oil and Gas Production Design Considerations

7,4.2.1. Inhibitor EfficiencyInhibitor efficiency is defined from laboratory measurements, as the relative corrosion rate with and without inhibitor:

C R -C RInhibitor efficiency =

CRx 100% (9)

where CRinh - corrosion rate in the presence of inhibitor and CR0 = corrosion rate in the absence of inhibitor.

The inhibitor efficiency is a function of inhibitor concentration and, is typically above 90% for successful inhibitors. This figure is often used in the determination of corrosion severity and the subsequent corrosion allowance. The system inhibitor efficiency w ill, of course, be influenced by the dispersion mechanism and, in particular, how the inhibitor partitions between the different phases present (Section 7.4.3). In this respect, it differs when measured in water alone or water/oil mixtures.

The calculation of corrosion allow ances or design lives generally starts by calculating the expected corrosion rate in the absence of inhibitor determined by the prediction models (Section 6.1) and adjusting factors outlined above in Section 6. The expected corrosion rate is then calculated by multiplying by (1 //100), where/ is the system inhibitor efficiency; the required corrosion allowance is then the design life times the inhibited corrosion rate.

Proposed values for the system inhibitor efficiency vary between 80 and 99%. The corrosion allowance (or design life for a given allowance) is very sensitive to the value of system inhibitor efficiency chosen; thus an efficiency of 85% will require 3 times the corrosion allowance of an efficiency of 95% and 15 times the corrosion allowance of an efficiency of 99%; an efficiency of 90% would require an allowance of 2 and 10 times, respectively.

At 99% the design lives are so large that the effects of temperature and C 0 2 partial pressure are negligible, i.e. when based on this quality of inhibition.

Until recently, most operators using this approach recommended a figure of 85% for design purposes. However, in light of past experience and recent advances in inhibitor performance, design figure have increased to 90%. Tests in the laboratory have given efficiencies well above 95% and it is felt that, given careful inhibitor selection, values of 90% can be achieved in the field in straight pipe under typical pipe wall shear stresses and in the absence of highly energetic flow (e.g. at tees and in slug flow conditions). This value is more in line with industry practice.

7.4.2.2. Inhibitor PerformanceIt has been frequently shown that the residual corrosion rate under inhibition does not display the same sensitivity to operational parameters as the corrosion rate without inhibitor. It results that the above mentioned approach cannot be but an approximation, especially if the prediction of CR0 itself is questionable. Therefore, some operators are directly selecting inhibitors according the resulting residual dissolution rate CRinh. As an example, the choice of chemical and dose rate must achieve a certain corrosion damage or rate.

In such an approach, the corrosion allowance is chosen first, on a technical and economical basis. This defines the mandatory inhibitor performance on the basis of

wh

7.4.1

par and are a sl con

Ccorican

T tl and corrc will great drop effici wate dosa; phase disru amer of ini- eddie reduc

7.4.3.. The v total signif (C0) ./ equal ratio ( water

1-A

tive corrosion

ons

(9)

rosion rate in

t, is typically .rmination of em inhibitor rism and, in sent (Section oil mixtures, lly starts by ? t ^ ;ned by > in fection 6. 100), where/ m the design

nd 99%. The sitive to the /ill require 3 hie corrosion n allowance

I CO, partial

gu reof 85% advances in e laboratory itor selection, ai pipe wall and in slug

ibition does rrosion rate >t be but an . Therefore, ng residual e rate must

chnical and the basis of

which, treatment conditions are then selected.

7.4.3. Inhibitor Partitioning and Persistency

7.43.1 . GeneralIn a multiphase flow, any chemical which is soluble in more than one phase

partitions between them according to the solubility equilibrium. This is true for gases, and especially CH4 and C 0 2; this is also true for organic corrosion inhibitors which are soluble to some extent in water and oil. Consequently, an inhibitor can adsorb on a surface if, and only if, it is present in the wetting (oil or water) phase in sufficient concentration.

On the other hand, the beneficial effect of a corrosion inhibitor is achieved in the corrosive phase only, which is the water phase. Two basic treatment philosophies can be established, respectively based on oil or water soluble products:

1 Oil soluble inhibitors (OS): display the best persistency in water, but they can be washed out if not replenished by a periodic wetting of the wall by oil.

2 Water soluble inhibitors (WS): have a weaker persistency. In case of high flow rates, they are also more sensitive to mechanical stripping, particularly if a high shear stress is applied.

The reduction in corrosion rate due to inhibition depends on a number of factors the basic efficiency of the inhibitor under standard conditions, the effect of flow and turbulence on the inhibitor efficiency and the dispersion of the inhibitor into the corrosive medium. For the oil phase, after the first stage separator, the inhibitor injected will need to be oil soluble to be dispersed throughout the oil phase. However, the greatest risk of corrosion will be at bends in the pipework where water is likely to drop out and the flow regimes at these locations may greatly reduce the inhibitor efficiency. In the produced water phase the inhibitor will have to be water soluble or water dispersible to ensure dispersion through the medium. If injected at a sufficient dosage, then this should be effective even at bends for moderate velocities. In the gas phase, inhibition is not likely to be effective, the high turbulence and flow rates will disrupt inhibitor films, hence no inhibited corrosion rates are presented for the amended corrosion rates for gas streams. In separators, there will again be the problem of inhibitor partitioning into the produced water phase from the oil phase. In addition, eddies and turbulent vortices set up by weir and vessel attachments will greatly reduce inhibitor efficiency

C 02 Corrosion Control 31

7 .4 3 2 . PartitioningThe injection rate of a commercial inhibitor is usually calculated according to the total liquid flow. It is then reported as an average content (C..; ,..;), whereas the significant concentrations are the respective concentrations in water (Ca,) and in oil (C }). At the solubility equilibrium, the activities of any chemical in water and oil are equal. Provided the activity coefficients do not change with concentrations, then the ratio of their concentrations is constant. This is the partitioning coefficient K between water and oil:

K = C:i./C (10)

For oil soluble inhibitors (OS), K is very low (e.g. < 0.1): these products 'do not pass' into water. For commonly available WS products, however, K is often balanced. It is thus necessary to make a distinction between 'water soluble' products (K = 1) and products with a "preferential solubility in water" (K 1).

For WS, the purpose is to actually get them into the water phase, whereas they are added with respect to the total liquid flow7. W'hence a first notion of 'partitioning efficiency' (PE), which is the ratio of what is obtained in the water to what is injected 157]:

32 CO, Corrosion Control in Oil and Gas Production Design Considerations

PE =C K

C , BSW ^^mcan 1+ ....... (K - 1)100

(11)

This efficiency may be >1.However, if the efficiency of a treatment depends on concentrations, its cost

depends on the corresponding quantities. Whence a second notion of 'economic efficiency' (EE), which is the ratio of what is useful to what is paid:

EE =C V BSW x P E

100(12)

where OS = oil soluble inhibitors, WS = water soluble inhibitors, PE = partitioning efficiency, EE = economic efficiency, CmaJn - average inhibitor concentration based on total fluids, C0 = concentration in oil, Cw - concentration in water, BSW - water cut (%), K = partition coefficient, Vw - volume of water, Viol = total volume, CRjnf - corrosion rate in the presence of inhibitor, CR0 ~ corrosion rate in the absence of inhibitor,, and / = inhibitor efficiency.

PE and EE are not only dependent on the partition coefficient K, but also on the BSW (Fig. 6). In particular, for traditional products with K < severe underdosage may be encountered at low BSW. Despite a water solubility, if its K is < 1, all the added product may then pass into oil, resulting in an unexpected depletion of C much below the 'expected' Cmegn. With the recent practice of transporting unprocessed crude oil, many of the recent inhibition failures are believed to be due to this unexpected underdosage at low water cuts.

Unfortunately, this convenient approach cannot be quantitatively applied. In fact, all the components of a commercial blend have their own partitioning coefficient. It follows that partitioning varies for one component to another. Consequently, what is usually called a partitioning effect is indeed a 'fractiorsing effect'. The usual presence of surfactants in a commercial blend also make the partition coefficient of all its components to vary with the added doses of the blend. This makes any modelling totally non-manageable.

The only way to take partitioning into account then is to replicate it experimentally. The fabrication of the test solution for electrochemical testing thus requires to stir strongly the real virgin oil and reconstituted water at the expected BSW, and then carefully to separate that water. During this extraction procedure, the hi situ pH

must least temp this s

7.43.. As al persi:

PvhP

Xfla

T h er

j- s n

par

C 02 Corrosion Control 33

its 'do not balanced, cts (K s 1)

is they are irtitioning is injected

(10)

(1 1 )

s, its cost economic

(12)

irtitioning ion based V = water ve, CRinh = ibsence of

Iso on the ierdosage : 1, all the don of Cw p;i ssed ue to this

^d. In fact, fficient It iy, what Is I presence t of all its nodelling

imentally. res to stir and then

n situ pH

must be truly reproduced. The corresponding acid gases must also be present, at least qualitatively. Temperature m ust be as close as possible to the expected temperature. This is not a problem as long as the latter remains below 60-80C. Above this some approximations cannot be avoided.

7.4.33. PersistencyAs already stated, this word covers two notions, namely physical and mechanical persistency.

Physical persistency expresses the ability of an adsorbed molecule to resist washing out in virgin water. It is measured by a 'time of defilming'. It is much higher for OS than WS, and it is improved by chem isorption and film polymerisation. It is a major feature for OS and batch treatments.

Mechanical persistency expresses the ability to resist a peeling effect by the fluid flow. It is traditionally measured in flow channels or on rotating devices,at increasing shear stresses or mass transfer coefficients, and on bare surfaces.

The relevance of this experimental method to field conditions is explained below.

High shear stresses are invariably present in the field, at least locally. With the success of inhibition, it can be postulated that high local flows in general do not necessarily prevent inhibitor effectiveness.

^w ateJ & mean

% water

Fig. 6 Influence of water content on the partitioning efficiency (C.wateJC ), for different partitioning coefficients (CwjlJ C Ml) [571.

Inhibition is an electrochemical process. It thus occurs at the metal-electrolyte interface, i.e. beneath any solid, but porous, surface layers (corrosion layers or original mill scale or rust layers). On the other hand, hydrodynamic effects only occur at the interface between a macroscopic solid and the bulk fluid, i.e. on the top of the solid surface layers. It follows that on scaled surfaces there is no shear stress applied where inhibition actually occurs, and that the mass transfer coefficient there is not flow dependent.

The loss of inhibition in the presence of high flow rates may resemble features similar to morphologies of erosion-corrosion. This has been interpreted as a flow-dependent loss of inhibition. However, this can also be due to erosion- corrosion after a totally flow-independent loss of inhibition, In view of this, field experience needs to be reconsidered.

As an example, oxygen entries can arise in the field, in the form of air entries into low pressure circuits, or in the recycling of aerated liquids (e.g. slops, sump caisson). This does change the performance of oil field corrosion inhibitors. For example, some may become slightly anodic, and overpass their critical pitting potential [57], thus explaining a loss of inhibition.

Similarly the existence of multiple steady states in CO., corrosion [531 is related to the irreversibility of surface states. This may also occur for inhibitor films, a subject not addressed to date and requiring further extensive research.

7.4.4. Commercial Inhibitor Packages

Commercial corrosion inhibitor packages contain active molecules, a solvent which is the carrier fluid, and additives, which are aimed at optimising the properties at all stages from delivery and storage to injection and production (e.g. co-solvents, surfactants, stabilisers, antifoaming agents, etc.). A 'commercial' inhibitor is thus a blend, whose activity is currently much higher than that of the non-formulated active

- molecule.Prior to being applied, an inhibitor must be proven to be an innocuous additive.

Therefore, its technological properties (non-foam ing, non-gunking, etc), often improperly called 'secondary', must be tested before the field application. Based on these requirements, an inhibitor package needs to incorporate a combination of characteristics to be effective.

34 CO, Corrosion Control in Oil and Gas Production Design Considerations

7.4.5. Inhibitor Compatibility

It should be noted that corrosion inhibitors can act as emulsifying agents and can lead to foaming which may decrease the effectiveness of the separation process. Similarly, emulsifying agents added to the process stream can interfere with effective inhibition. There may, in addition, be compatibility problems between the corrosion inhibitor best suited to the oil phase and the corrosion inhibitor best suited to the produced water. Finally, the presence of thick scales may reduce the effective inhibitor efficiency, but increase inhibitor performance!

Sin ce i transm Three r

dis me

tra

tra

Corrpossibl

TheSection

Downlwith h) such as inhibitc to inject stream behind exacert

Ano gas lift vaporis difficul

This

Dry gainhibiti pipelin adequa at appr

Crude n layers ic effects Tuid, i.e. 3 there is he mass

features 2ted as a srosion- / of this,

entries into ip caisson), nple, some 1 [57], thus

s related to s, a subject

vent which erties at all o-solvents, :>r is thus a ated active

is

36 C 0 2 Corrosion Control in Oil and Gat Production Design Considerations

7.4.7. Inhibitor Distribution in Multiphase Pipelines

Because of transient behaviour in multiphase flow, both inhibitor transport and inhibitor dilution effects are important parameters.

The transportation time of inhibitors needs to be known for proper inhibition. For example, different phases move at different velocities along a pipeline. The liquid residence times and consequently the time to transport inhibitor to the desired location, can be very long. At low demand conditions, or towards the end of field life, a gas or liquid pipeline may be greatly oversized and times to transport corrosion inhibitor from the inlet to the outlet may be very long, even exceeding a few months.

Inhibitor distribution along the top of the line is limited in stratified flow. Furthermore, inhibition of fluids flowing in the slug flow* conditions poses a major problem and is less effective.

Fluid flow will affect the deposition and movement of sand and other debris in the flowline. The design must consider inhibition under the various flow conditions and determine whether movement or settling of sand will have an impact on inhibitor performance.

Inhibitor dilution can occur in multiphase (especially wet gas) pipeline sections due to both hydrocarbon or water condensation. The inhibitor concentration may be reduced significantly from the concentration determined at average inlet flow conditions.

7.4.8. Effect of Flow on Inhibition

Flow conditions can influence inhibitor effectiveness in a variety of ways. Inhibitor transport to the pipe surface will depend upon the flow regime while adsorption / desorption processes will be affected by the local fluid conditions or fluid-to-wall shear stress.

Pipelines operating under annular, mist or dispersed flow regimes can be inhibited relatively easily because the mixing of the turbulent gas and liquid phases provides good contact of inhibitor with the pipe wall. Effective inhibition is more difficult under stratified flow where there are two main areas of concern. Firstly, at low spots in the line where stagnant liquid may collect and the local turbulence may not be sufficient to allow adequate mixing of inhibitor. The second problem area is on the top of the line where fresh and renewed condensing water is corrosive. In this case, vapour transport of inhibitor to the top of the line is required. Alternatively, frequent batch treatment by an inhibitor slug is needed.

The removal of inhibitor films is generally worse w