Chapter 2

Corrosion basics and computer modeling

V.G. DeGiorgi
Naval Research Laboratory, Department of the Navy, Washington
D. C. 203755000, U.S.A.

1. A layman’s overview of corrosion

1.1 The cost of corrosion

Corrosion is the destruction of material resulting from exposure and interaction with
the environment. While common usage typically associates corrosion with metals, the
destruction of non-metallic materials as a result of exposure to the environment can
be considered corrosion; the disintegration of plastics when exposed to sunlight and
the rotting of wood are examples of non-metallic corrosion. However, the concern
of this chapter is corrosion of metals, especially electrochemical corrosion, and the
ability to use computer modeling techniques to predict corrosion.
Corrosion is a natural phenomenon which exists as part of our everyday life. In
extreme situations catastrophic failures such as sudden collapse can result when corro-
sion caused damage undermines the load bearing capability of a structure. Corrosion
damage can also result in life threatening situations which are not as dramatic as
structural collapse such as contamination by corrosion products from lead piping re-
sulting in possible toxic levels of lead in drinking water supplies.
The cost of corrosion damage is felt by industry, government and individuals re-
gardless of gross earnings, political system or life styles. Repair or replacement costs
for bridges and other highway structures are a major cost incurred by regional govern-
ments. Repair and replacement costs are incurred by any industry which uses metal
structures in processing or storage facilities. Individuals are subject to direct costs
related to repair or replacement of personal property, such as automobiles or home
hot water heaters, and to indirect costs in the form of higher taxes or higher product
prices which result from government and industry costs.
Corrosion is a major problem which must be addressed for safety, environment
and economic reasons. Safety concerns are paramount in any situation. Environmen-
,.
48 Industrial Applications of the BEM

tal concerns include consideration of corrosion caused pollution, depletion of resources

such as those needed for replacement of the corroded structure and disposal of the
corroded structure. Safety and environment al concerns tend to be very hard to define
in terms of costs. Economic concerns, however, tend to lend themselves readily to
cost estimates. For example, the total cost of metallic corrosion in the United States
for 1975 was estimated at $82 billion or 4.9% of the Gross National Productl. This
estimate of corrosion considered preventable and non-preventable corrosion. Since
corrosion is an inescapable natural phenomenon, it may be more appropriate to esti-
mate only that portion of corrosion damage which cannot be reduced through design
or application of preventative technology. Non-preventable corrosion accounted for
60% of the total which corresponds to a cost of $49 billion for the year. Another
method for estimating the cost of corrosion is to consider the impact on a particular
community. For example, in the paper industry it was estimated that the cost of cor-
rosion damage increases the cost of paper between $6 and $7 per tonz. Still another
method for estimating the cost of corrosion is to determine the percentage of yearly
output which is loss to corrosion. For example, approximately a third of all metal
produced by the Soviet Union steel industry in 1965 was loss to corrosion. Again,
the issue of what costs should be considered when calculating the cost of corrosion
damage must be considered. The amount of metal loss to corrosion, after the recovery
of scrap metal is deducted from the total metal loss was 10% of the yearly metal out-
put. As can be seen from these three methods of estimating a cost of corrosion, the
economic impact of corrosion is immense but not as well defined as one may originally
believe. It is obvious, however, that corrosion is a major problem with large financial
implications which must be addressed.
The study of corrosion involves the study of the chemical, physical, metallurgical
and mechanical properties of materials. Corrosion is a synergistic phenomenon in
which the environment is equally as important as the materials involved. Computer
modeling techniques are a powerful tool for the study of complex system response
and are therefore an appropriate means for studying corrosion. The theoretical basis
needed for the application of computer modeling to selected corrosion problems has
existed for several decades4. The pace of adaptation of computer modeling techniques
to corrosion evaluation has been set by the availability of computer resources and
the cost of necessary computer power rather than the development of the mathemat-
ical representation of corrosion. The cost and availability of computer resources is
constantly changing. Advances in computer technology have tended to result in a
continual decrease in computational costs which in turn makes computer modeling a
more viable option.
In the recent past computer modeling techniques have been successfully applied
to corrosion problems as summarized in review articles by Zamani, Porter and Mufti5
and Munn6. The application of computer modeling techniques to corrosion systems re-
quires an understanding of the physical phenomenon of corrosion and the mathematics
which govern the corrosion process. In addition, a knowledge of the numerical proce-
dures which are the basis of computer modeling techniques is necessary for accurate
computational analyses. Verification of assumptions by comparison of computer anal-
ysis results with experimental or other measured data is a fundamental requirement
in the creation of an accurate analysis. The purpose of this chapter is to introduce
 Corrosion basics and computer modeling 49

the individual who is knowledgeable in computer modeling techniques to the basic

concepts of corrosion. Basic corrosion terminology and an introduction to corrosion
processes are presented. Detailed examination of the theoretical and applied aspects
of corrosion can be found in textbooks, study guides and technical journals. Examples
of corrosion evaluation using boundary element techniques are presented in the final
section of the chapter.

1.2 Chemical and electrochemical corrosion

Corrosion is either chemical or electrochemical in nature. The distinction between

chemical and electrochemical corrosion is based on the corrosion causing mechanism.
Chemical and electrochemical corrosion are not mutually exclusive and can occur si-
mult aneously.
Chemical corrosion is the direct result of exposure of a material to a chemical
and is governed by the kinetics of chemical reactions. Chemical corrosion does not
involve the generation of an electrical current. Direct chemical attack, such as the
dissolutive of a material by an acid, and selective attack, such as leaching of a specific
soluble compound from a material, are two common forms of chemical corrosion. The
formation of oxide films on a metal surface is also a form of chemical corrosion.
Electrochemical corrosion is the dissolution of a metal through the oxidation pro-
cess. Oxidation and reduction chemical reactions occur simultaneously and are inter-
dependent. Corrosion only occurs at the site of the oxidation reactions. Oxidation
involves the loss of electrons; reduction involves the gain of electrons. The electron
transfer between oxidation and reduction reaction sites establishes the electrical cur-
rent required for electrochemical corrosion. There is no gain or loss of electrical charge
during electrochemical corrosion.
The basic unit in which electrochemical reactions occur is the electrolytic cell. An
electrolytic cell consists of three components; an anode, a cathode and the electrolyte.
The anode is the site of oxidation and therefore the site of corrosion. The anode and
cathode are electrically connected. The medium surrounding the anode and cathode
is the electrolyte. The anode and cathode can be the same or different metals. When
a structure fabricated of one material is surrounded by an electrolyte and exposed to
an electrical current, portions of the structure will act as anodes while the remainder
of the structure will act as the cathode. When a structure is fabricated of two metals,
the more likely to corrode, or the more active, metal will act as the anode while the
less likely to corrode, or nobler, metal will act as the cathode. The metallurgical
properties of the anode and cathode directly effect the rate and extent of corrosion.
The relative surface area of the anode and cathode will effect the rate of corrosion.
Large cathode areas provide a large surface area for reduction to occur. The rate of
oxidation of the anode is regulated, in the absence of other limiting factors, by the
number of electrons required for reduction of the cathodic surface. A small anodic
area will require a greater rate of oxidation to supply the same number of electrons
as a larger anodic area. Therefore, the smaller the anodic area with respect to the
cathodic area, the greater the rate of corrosion on the anodic area. Conversely, the
rate of corrosion can be reduced by increasing the relative size of the anode.
An electrolyte may be a fluid, a gas, a granular solid such as sand or a more com-
50 Industrial Applications of the BEM

plex solid and fluid combination such as a saturated or partially saturated soil. The
composition and conducting properties of the electrolyte effect the rate and extent of
corrosion. Environmental conditions such as temperature, humidity for gases, water
content for soils and pressure effect the rate and extent of corrosion.
When a metal is surrounded by an electrolyte, the metal will establish an electri-
cal potential with the surrounding medium. Electrical potential is directly related to
the electrical current in the material. The open circuit or rest potential is the mea-
sured electrical potential which occurs when there is no flow of electrical current in
the circuit formed by the electrolytic cell. Measured electrical potential is a potential
difference. The measured electrical potential is the difference between the electrical
potential at the point of interest and the potential of a reference electrode. Reference
electrodes are usually chosen based on availability, survivability in the working envi-
ronment, and traditional use in the working environment. Corrosion is often reported
in terms of weight loss per unit time or thickness loss per unit time. The corrosion rate
is directly proportional to the measured electrical current. The mass loss is related
to the measured electrical current by Faraday’s Law:

nFW
Q=T (1)

where

F = Faraday’s number (96 500 Coulombs/equivalent),

~ = the number of electrons transferred
W= weight of reacting species
M= molecular weight of reacting species
Q = electrical charge = It
1 = electrical current
t = time

If corrosion is assumed to occur uniformly over the metal surface, the corrosion rate
can be directly calculated from the mass of metal loss due to corrosion. Rearranging
Eq. (l):
(2)

Substituting Q = It into Eq. (2):

(3)

Final re-arrangement yields weight loss per unit of time:

W IA4
—=— (4)
t nF
1.3 Galvanic corrosion

Galvanic corrosion occurs when two or more metals with different electrochemical re-
sponses are electrically connected. In the absence of an externally applied electrical
current, the difference in open circuit potential creates the electrical current necessary
 Corrosion basics and computer modeJing 51

for electrochemical corrosion. It impossible tousethe differences in open circuit po-

tentials to predict corroding tendencies; galvanic series, such as that shown in Table 1,
indicate the relative tendency for metals and metal alloys to corrode. The relative
position of the metal in a galvanic series, rather than the value of electrical potential
associated with the metal, is indicative of the tendency for corrosion to occur. For
example, the more noble or cathodic metals such as platinum, gold and titanium show
little tendency to corrode while the more active or anodic metals such as zinc and
magnesium alloys will corrode when coupled wit h any more noble metal. Some metals
appear twice in a galvanic series; steel on which a protective oxide film has formed
is more noble than the same steel without the protective oxide film. The closer ma-
terials are to each other in the galvanic series, the less galvanic corrosion will occur
when the two materials are used in combination. When test data for a specific mate-
rial pairing does not exist, the galvanic series for the environment in which the couple
is immersed can be used as an indication of the degree of severity of galvanic corrosion.

1.4 Physical varieties of corrosion

Corrosion damage can be divided into many categories based on the physical char-
acteristics of the damaged metal. Simple definitions of a few frequently encountered
types of corrosion damage are given in this section. This abbreviated list of physical
corrosion damage categories is meant to give the reader a glimpse at the variety of
corrosion related damage possible.
General corrosion gives the appearance of an uniform wasting away of the metal.
General corrosion is often the result of multiple adj scent anodic and cathodic regions
on the metal surface. The corrosion appears to be uniform over the metal surface
due to the relatively small spacing between anodic sites. Detailed examination of
the apparently uniformly corroding surface on a microscopic scale will find a mixed
pattern of anodic and cathodic regions.
Cavitation damage occurs at solid-liquid interfaces, such as the surface of a sub-
merged structure. Repeated collapse of vapor bubbles may cause physical damage to
the solid surface which may result in pits. Pits can also be formed by preferential
corrosion at sites where the collapse of vapor bubbles has resulted in localized damage
of oxide films. The relatively small areas of film damage result in small localized an-
odic regions which experience accelerated corrosion due to a large cathode to anode
surface area ratio. Pits are formed by corrosion which extends below the original
metal surface.
Fretting corrosion is localized damage between contacting surfaces. The corro-
sion process is accelerated by relative motion of the two surfaces. Fretting corrosion
requires initial contact between the two metal surfaces as well as relative motion suf-
ficient to expose the surfaces to the corroding environment. Abrasion between the
surfaces, either due to grinding between the surfaces or scratches caused by grit or
other contamination between the surfaces, can result in damage to the metal surfaces
and to any covering oxide film. The damaged areas will then act as anodic sites for
accelerated corrosion due to large cathode to anode surface area ratios.
In microbiological induced corrosion the destruction of metal is caused directly
or indirectly by bacteria, fungi or other microbiological organisms. Anaerobic organ-
52 Industrial Applications of the BEM

isms, aerobic organisms and their digestive by-products all act as corroding agents.
Microbiological induced corrosion is often a time dependent system which involves
the layering of anaerobic and aerobic organisms. Some environmental characteristics,
such as temperature and flow conditions, can promote growth of the biological or-
ganisms involved and therefore enhance corrosion. Microbiological induced corrosion
can be controlled by the use of growth inhibitors such as chemical treatments and
stringent water quality cent rols.
Selective attack, or leaching, is the removal of a soluble compound from an alloy.
Dealloying of brass, or dezincification, is a common form of leaching. Dezincification
occurs when brass is electrically connected to zinc. The material left after dezincifi-
cation is a very porous copper with poor mechanical properties. The visible sign of
dezincification is a change from the typical yellowish color of brass to the reddish color
of copper. The copper material which remains is subject to sudden failures because of
the low material strength. Dezincification is only one example of leaching corrosion.
Stray current corrosion results from the flow of electrical current through a path
other than intended. The sources of stray current are often extraneous to the structure
damaged and unknown to the analyst until corrosion is observed. Even after a cor-
rosion problem has been identified, the source of stray electrical current may remain
a mystery. The stray current creates an anodic site at the point where it enters the
structure. The effect of the stray current may be to change a surface which is initially
a cathode into an anode. The resulting electrical circuit may have no resemblance
to the original circuit. Stray current corrosion often occurs due to faulty design of a
system other than the one in which corrosion is observed.

2 Material polarization response

In the evaluation of material or system performance an appropriate material charac-

terization is necessary. In the case of electrochemical corrosion the material behavior
of interest is the electrical current-electrical potential relationship.
The sensitivity of the electrical current-electrical potential relationship to envi-
ronmental and electrolyte characteristics must be considered when defining an ap-
propriate characterization. The accuracy of any performance or corrosion damage
prediction, whether based on analytical, experimental or computational analysis will
depend on how well the system defined in the analysis matches both the real structure
and the environment surrounding the structure. Factors which are essential in defining
the electrolytic cell where corrosion occurs are the geometry of the structure under
evaluation, the location of electrical sources or sinks, the strength of the electrical
sources or sinks and the characteristics of the electrolyte surrounding the structure.
Time dependent variation in electrical current-electrical potential response are impor-
tant for many situations such as the evaluation of structures which experience extreme
seasonal variations in temperature or water salinity. The accuracy of any evaluation
will depend on the accuracy of the system model and the electrical current-electrical
potential response. The number and magnitude of differences between experimental
and service conditions will determine whether analysis results will be approximate to
the real world situation or completely erroneous.
 Corrosion basics and computer modeling 53

2.1 Polarization definitions

Polarization is the relationship between electrical current and electrical potential. It

is an observed behavior resulting from the combined effects of multiple oxidation and
reduction react ions which are dependent on many variables. The relationship between
electrical current and electrical potential may or may not be linear. It may or may
not be possible to develop a mathematical function which represents the relationship.
Polarization is generally referred to as a characteristic of the metal but is dependent
on metal and electrolyte composition.
Anodic polarization response is the behavior of the corroding metal. Cathodic
polarization response is the behavior of the non-corroding metal. Much of the ap-
plication of computer modeling techniques to corroding systems is in the evaluation
of corrosion prevention systems. Corrosion prevention systems attempt to maintain
the metal of interest in the cathodic state so an understanding of the behavior of
non-corroding met al is essential.
Polarization response is either activation or concentration controlled. Activation
polarization occurs when the electrical current-electrical potential relationship is con-
trolled by the rate of a specific chemical reaction in the corrosion process. The only
limitation imposed on the critical reaction is its associated rate of reaction. Idealized
cathodic activation polarization, Fig. 1, is:

qact = p log $to (5)

where

q~~t = measured potential

/3 = constant (Tafel slope)
i. = initial electrical current
i= electrical current

Anodic response will follow the same mathematical form. Because activation polar-
ization is limited by the kinetics of a specific reaction, the polarization response is
dependent on temperature and all other parameters which effect the chemical reaction
rate.
Concentration polarization occurs when corrosion is limited by whether a specific
chemical reaction can occur. In other words, polarization is controlled by the con-
centration of a chemical component required by a specific reaction in the corrosion
process. One particular type of concentration polarization is diffusion controlled po-
larization. In this case, the ion diffusion rate to the metal surface limits the rate of
the reduction reaction since the diffusion rate limits the available of the required ion.
Idealized diffusion rate limited cathodic concentration polarization response (Fig. 1)
is:
RTa
7/ . . . = 2.3#og 1 – : (6)
( tL‘)
where
54 Industrial Applications of the BEM

qconc = me-red potential

R = gas constant
n = number of electrons transferred
T. = absolute temperature
F = Faraday’s number
iL = limiting diffusion electrical current
i= electrical current

Eq. (6) is based on the Nernst equation.

In summary, a limiting rate of a chemical reaction governs activation polarization
while the availability of a necessary chemical components governs concentration po-
larizat ion.
The total, or measured, polarization is a combination of activation and concen-
tration polarization. Activation polarization usually controls for corrosion related
reactions which have low chemical reaction rates while concentration polarization will
become dominate when the governing reactions have higher reaction rates. Measured
polarization response can be idealized by the linear combination of concentration and
activation polarization responses for cathodic polarization:

q. = Vact + qconc (7)

An idealized measured cathodic polarization curve is shown in Fig. 2.

Linear polarization is, as the name implies, that portion of the total polarization
response which is linear in nature with respect the logarithm of the electrical current
density. The slope of the linear portion of the polarization curve is often referred to
as the Tafel slope.
The effects of the formation of oxide films are not addressed by the polarization
curves presented in Figs. 1 and 2. The formation of oxide films greatly reduces the
electrical current associated with a defined value of electrical potential and the po-
larization relationship becomes multi-valued with respect to electrical current. The
value of electrical potential associated with an electrical current value will depend on
whether the reading is taken prior to oxide film formation, after film formation or after
film destruction. The effects of film formation is demonstrated for polarization which
is initially anodic in Fig. 3. The initial behavior of the metal can be described using
activation polarization definition as given in Eq. (5). Prior to oxide formation the
metal represented in Fig. 3 is exhibiting anodic behavior. The formation of an oxide
film transforms the active material to a more noble material with a reduced tendency
to corrode and the electrical current density is reduced to a relatively small limiting
value. The limit ing value of electrical current density may be as much as 106 times
lower than the maximum active region electrical current density7. At higher levels of
electrical potential, oxide films tend to break down and the material is returned to
an active state showing increased anodic polarization; this is the transpassive region.
The formation of oxide films may cause a metal to shift from anodic to cathodic and
finally return to anodic behavior depending on the other metallic components of the
electrolytic cell.
 Corrosion basics and computer modeling 55

2.2 Mixed potential response

When two metals are electrically connected and surrounded by an electrolyte, the
measured elect rical potential of both metals will change from the open circuit poten-
tials. The cathodic metal will become more anodic and the anodic metal will become
more cat hodic. The two metals will reach electrical potentials which are somewhere
between the open circuit potentials of the two metals.
The mixed potential theory defines this electrochemical corrosion behavior. The
hypotheses of the mixed potential theory are:

- Any electrochemical reaction can be divided into multiple partial oxidation and
reduction reactions.

- There is no accumulation or loss of electrical charge during the electrochemical

corrosion process.

The first hypothesis recognizes the chemical kinetics of electrochemical corrosion. The
second hypothesis requires that the oxidation and reduction reactions are in a state
of equilibrium. It is this balancing of reactions which causes the anodic and cathodic
metals to achieve some intermediate electrical potential when connected. The mixed
potential theory is one of the basic tools for the evaluation of electrochemical corro-
sion. It is important to note that the concept of mixed potential is a theory which is
based on experimental observations and correlation between theoretical and experi-
mental response should not be taken as proof of the theory8. This does not diminish
the importance of the theory.

2.3 Electrolyte characteristics

The amount of corrosion which occurs in any system is dependent on both the metal
and the electrolyte. The corrosion response of a metal, represented by the measured
polarization response, will vary depending on the constituents of the electrolyte and
their concentrations. In establishing experimental criteria to determine polarization
response it is best to match experimental and service condition electrolytes as near
as possible. In order to introduce the complexities encountered due to electrolyte
constituents two characteristics of fluids, the hydrogen ion concentration and the dis-
solved salt concentration, are briefly discussed.
Reviewing basic chemistry, the hydrogen ion concentration is measured by the pH
of a solution. Metals which are stable in both acidic (pH < 7.0) and basic (pH > 7.0)
solutions exhibit corrosion damage which is independent of hydrogen ion concentra-
tion. Some metals, such as zinc and aluminum, exhibit a parabolic corrosion-hydrogen
ion concentration behavior; corrosion rates are high for high and low pfl values while
corrosion rates are near zero for neutral and low pH values in both the acid and basic
range. More typical of metals is a direct relationship between rates of corrosion and
level of acidity; the more acid the solution, the higher the rate of corrosion.
Dissolved salt concentration effects on the electrolyte-material system are depen-
dent on the concentration and type of dissolved salts. In general, there are aggressive
salt ions and inhibitive salt ions. Aggressive ions accelerate corrosion activity when
compared to a salt free solution. In general inhibitive ions decrease the corrosion rate
56 Industrial Applications of the BEM

through the formation of protective films. Ions can be either aggressive and inhibitive
depending on their concentration and the concentration of oxidizing agents in the
electrolyte solution. Aggressive salts ions, such as Cl–associated with seawater, will
become corrosion inhibitors if there is insufficient oxygen available. This effect of oxy-
gen content can be seen in the observed corrosion rates for metals exposed to seawater
environments; maximum corrosion occurs in the splash zone where oxygen is readily
available and minimum corrosion occurs at depths where only a limited amount of
oxygen is available.
Electrolytes may be homogeneous or non-homogenous. In order to determine if
an electrolyte can be considered as a homogeneous medium, the electrical behavior
of the volume in the immediate vicinity of the anode and cathode must be evalu-
ated. Variations of electrolyte constituent near the anode and cathode surfaces will
have a greater effect than variations at a distance. At distances sufficiently far from
the anode and cathode surfaces, variations in the electrolyte constituency will have
negligible effects on the measured polarization response. The volume of electrolyte
which has an influence on measured polarization response will vary depending on
the materials involved and the electrolyte. An electrolyte may be considered to be
homogeneous if there are no local variations which result in variations in material po-
larization response. For example, high levels of turbulence near a metal surface may
result in localized variations in polarization response. However, in the absence of high
flow rates or turbulence a complex mixture of dissolved salts, biological agents and
other constituents such as seawater can be considered to behave as a homogeneous
electrolyte.

2.4. Experimental polarization curves

Measured polarization response, determined through appropriate experimental pro-

cedures, incorporates all effects of activation polarization, concentration polarization,
formation or destruction of oxide films, formation of any biological layers, effects of
electrolyte characteristics and oxidizing agent availability. Briefly stated, the mea-
sured response is a synergistic total response which explicitly includes all factors.
As in any experimental procedure, there are minimum requirements for the successful
measuring of polarization response. A primary consideration in designing experiments
to measure the polarization response is to design the experiments so the primary cor-
rosion mechanism is electrochemical corrosion. Monitoring and cent rolling methods
must be established which do not disturb the system in ways which result in mean-
ingless measurements; this issue is common to any experimental procedure.
In an ideal situation, special test procedures to determine the polarization re-
sponse would be defined which correspond to expected service conditions. Special
test specimens which are representative of the structure to be evaluated would be
used in the experiments. Unfortunately time and financial considerations may elim-
inate all possibilities of special testing procedures. Therefore, a basic understanding
of the standardized test methods which may be used to obtain polarization response
is necessary.
Polarization response testing can be divided into two broad subgroups; constant
electrical current and constant electrical potential tests. In constant electrical current
I
Corrosion basics and computer modeling 57

experiments the polarization curve is developed from measured potential values for
defined values of electrical current. Discrete steps are taken between different electri- I
cal current levels. The assumption of constant electrical current testing is that the
polarization response is a single valued function with respect to electrical current.
With the formation of oxide films, polarization response does not maintain this sin-
gle value characteristic. Therefore, constant electrical current, or galvanostatic, tests
methods are not capable of defining changes in polarization response which result
from the formation of oxide films. Galvanostatic tests are therefore of limited use for
oxide film forming metals which will be in service for any significant period of time.
Potentiostatic and potentiodynamic testing are based on maintaining a constant value
of electrical potential. These testing techniques can be used to obtain valid polariza-
tion response curves for metals which form oxide films.
Potentiostatic testing involves the adjusting of electrical current to maintain a con-
stant electrical potential value. The polarization response is developed by creating a
series of electrical current-electrical potential data points. The changes in electrical
potential are defined as part of the experimental procedure. The input electrical cur-
rent is increased until the desired electrical potential is reached. Since potentiostatic
conditions allow the establishment of steady state conditions at each electrical poten-
tial level, the test method allows for the formation and growth of protective films. In
potentiodynamic testing procedures the electrical potential is continuously increased.
Electrical current and electrical potential are recorded for specific intervals in time.
These two test methods should produce interchangeable resultsl” but there is evidence
that this is not a valid assumption for some materialsll.
The choice of experimental technique to determine polarization response should
be based on the expected service conditions. Ideally, the exact service conditions
are duplicated in the experimental determination of polarization response. However,
when this is not possible, test procedures which most closely resemble the service
conditions should be selected.

3 Mathematical basics

In electrochemical corrosion there is no loss or gain of electrical current during the

where i ., iv and i= are the components of the electrical current vector.

The electrical current is directly related to the first derivative of the electrical I

where a is the conductivity of the electrolyte.

Combination of Eqs. (8) and (9):

(lo)
58 Industrial Applications of the BEM

The conductivity may be a functional relationship or a constant value.

Assuming a constant value for conductivity allows for the reduction of Eq. (10)
to the Laplace equation:
i32@ + a2@ + 820
~ ~ ~=o (11)

When the Laplace equation is used as the governing equation for electrochemical cor-
rosion the effects of electrical sources and sinks are incorporated into the mathematical
model through boundary condition definitions.
Regions of the boundary which have fixed values of potential are defined by:

@=cl (12)

where (2’1 is a constant. Voltage sources are boundary conditions of the first or Dirich-
let type. In this type of boundary conditions the dependent variable, the electrical
potential, has a defined value. Defined values of voltage can be used to model im-
pressed anodes or other electrical sources where the voltage is held constant for steady
state conditions.
Electrical current sources or sinks are defined by:

(13)

where n is the normal to the boundary and L’z is a constant. Electrical sources and
sinks are boundary conditions of the second or Neumann types. The normal derivative
of the dependent variable, the electrical current or current density, is assigned a defined
value. Painted surfaces are oft en represented as:

(14)

Neumann boundary conditions can also be used to model impressed current anodes
or other electrical current sources when the electrical current can be represented as a
constant value. Any electrical source which does not vary in time can be represented
in this fashion.
The electrical current, or current density, of an exposed metal surface is deter-
mined from the material polarization response and the measured electrical potential
on the metal surface:
i = ~(~) (15)
where ~ represents a functional relationship. In the Tafel response region, the po-
larization relationship between electrical current and electrical potential is linear,
however, in general the relationship is nonlinear. The anodic and cathodic polariza-
tion responses typically are not represented by one analytical solution. Polarization
response is a boundary conditions of the third or Cauchy type in which the boundary
conditions has a functional dependency on the dependent variable, the electrical po-
tential in this case. Cauchy boundary condition are also known as mixed boundary
conditions.
 Corrosion basics and computer modeling 59
,P
/
4 Corrosion prevention

Corrosion is a natural phenomenon which cannot, using strict definitions, be elimi-

nated. However, if the rate of corrosion can be reduced so that the amount of corrosion
which occurs over the lifetime of a structure is negligible, corrosion has been, for all
practical purposes, eliminated. Good design practices and a knowledge of fundamen-
tal corrosion behavior can be combined to greatly reduce corrosion damage. Often
the primary goal of computer modeling is to evaluate the performance of a corrosion
prevention system.
Three common corrosion prevention techniques are summarized in this section;
material selection based on electrical compatibility, cathodic protection and painting
or coating. The techniques described are not the only methods available to reduce
corrosion. Each of the techniques can be readily modeling with existing boundary
element techniques.

4.1 Material selection based on electrical compatibility

Galvanic corrosion occurs when materials having different open circuit potentials and
polarization responses are electrically connected. The amount of galvanic corrosion
which occurs can be reduced by judicial selection of materials based on their location
in the galvanic series or by more detailed evaluation of the electrical current generated
when the two materials are connected. The magnitude of the electrical current will
depend on the electrolyte conductivity, the ratio of cathodic to anodic material surface
area, and the cathodic and anodic material polarization behavior. All of these factors
will determine the driving force for the resulting electrical current. The greater the
electrical current generated, the greater the corrosion rate on the anodic material.
In the design of joints or connections in which two or more different materials are
used, galvanic corrosion response must be taken into consideration. Material compat-
ibility with respect to rest potential and possible corrosion damage is as important as
basic structural characteristics such as yield strength and ultimate strength.
Computer modeling techniques can be readily used to determine the galvanic cor-
rosion electrical current associated with different material combinations. Parametric
studies can be performed to determine which, of all possible material combinations,
results in the lowest galvanic electrical current and therefore the lowest corrosion rate.
Computer models will explicitly address geometric and relative area concerns as well
as determining the effects of differences in polarization responses.

4.2 Cathodic protection

Cathodic protection uses electrochemical properties of metals to insure that the metal
of interest becomes the cathode of an electrolytic cell. Corrosion is allowed to occur
but the electrochemical properties are taken advantage of in such a way that the an-
ode is defined and corrosion damage is contained to a specific area.
In galvanic corrosion, the more active metal preferentially experiences corrosion.
Cathodic protection using sacrificial anodes takes advantage of the galvanic properties
of two metals. The sacrificial anode is a volume of metal added to the structure for
60 Industrial Applications of the BEM

the purpose of being destroyed by corrosion. The sacrificial anode metal is chosen so
that it is more active than the metal to be protected. Classical examples of sacrificial
anode systems are zinc, aluminium and magnesium anodes used to protect steel struc-
tures in aquatic or moist environments, such as seawater spray, partially saturated
soils and saturated soils. Zinc, aluminium or magnesium are more active than steel so
these materials will become the anode and will preferentially corrode before corrosion
occurs on the cathode, i.e. the steel structure. Since the system is designed so that
the anode is known, the length of time required for complete destruction of the anode
by corrosion can be determined. Prior to complete destruction of the sacrificial an-
ode, additional sacrificial anodes can be placed on the structure providing a continual
supply of material which can corrode without effecting structural integrity.
Sacrificial anode cathodic protect ions systems are not perfect. Sacrificial anodes
may corrode non-uniformly and are not be able to adjust dynamically to increased
electrical current demands which result from changes in the service conditions, envi-
ronment or increased area to be protected due to surface damage. The structure itself
is required to have an increased load bearing capacity because of the added weight
of the sacrificial anodes. Uneven corrosion of the anodes will also effect the system
corrosion prevention performance. The region of cathodic material which can be pro-
tected by a single anode is dependent on the metal’s polarization characteristics and
the conductivity of the electrolyte. Despite the limitations, sacrificial anode systems
are widely used.
Existing boundary element methods, which incorporate first, second and third
type boundary conditions capabilities, can be used for modeling sacrificial anodes.
Sacrificial anodes can be defined as sources with physical dimensions and location or
can be defined as point sources. Prescribed voltage values along with defined polar-
ization response completely define the anode material behavior.
Impressed current cathodic systems and sacrificial anode systems prevent corro-
sion in the same manner; by catholically polarizing the material to be protected.
Impressed current systems use an external power supply as the source of electrons
required for polarization of the metal to be protected. An external power supply may
have more than one location, or impressed current anode, where current is supplied
to the system.
Some problems associated with sacrificial anodes are eliminated by the use of im-
pressed current systems. Impressed current systems can be adjusted to meet dynamic
changes in the system. Whether an impressed current system can adjust to provide
adequate protection will depend on the system design. The added weight of sacrificial
anodes is replaced by the added weight of the required power supplies. For a large
structure, the difference in added weight can be significant. The area which can be
protected by an impressed current anode is typically greater than the area which can
be protected by a sacrificial anode. The extent of metal which may be protected by a
single impressed current anode is dependent on the electrolyte conductivity and the
available power supply. The cathodic material polarization response, of course, also
plays a role in the extent of cathodic material which can be protected by a single
anode.
Impressed current systems may be modeled using existing boundary element tech-
niques. The impressed current anodes can be defined with specific areas and locating.
 Corrosion basics and computer modeling 61

They can be assigned values of electrical voltage or electrical current density. The
boundary conditions defining impressed current anodes are held constant through out
the analysis thus modeling a constant power supply.

4.3 Paints and coatings

Paints and coatings are used to protect from corrosion in the same manner; the paint
or coating is used to isolate the metal to be protected from the environment. A coat-
ing can be a layer of a non-corroding metal plated to the metal to be protected, a
paint and primer system or a naturally forming surface film. The term ‘paint’ may
or may not imply cosmetic or ascetic purposes in addition to corrosion protection.
Paints are a combination of pigment and medium containing the pigment. In this
brief description of paint and coating systems, the terms paint and coating will be
used interchangeably.
There are a wide variety of paints available and marketed for use as corrosion
inhibitors. Different paints have been developed for different materials and environ-
ments. The selection of a suitable paint is only the first step in the creation of an
effective corrosion prevention system. Surface preparation, application methods, num-
ber of layers, combinations of different layers of different types of paints or primers
and thickness of layers are features which define a paint system.
Unfortunately, painted surfaces are not free from defects regardless of the care
taken during application. Mechanical damage, such as scratches or chipping, and
flaws from incorrect application procedures result in areas of exposed metal. These
exposed areas can be either small or large. If only relatively small areas of metal are
exposed through paint damage, corrosion can become severe. The remaining painted
surface acts as the cathode while the small exposed region of metal acts as the an-
ode. The large cathode to anode surface area ratio results in severe corrosion attack
because of galvanic corrosion behavior. In order to avoid severe corrosion damage at
small isolated areas of paint damage, secondary protective systems, such as cathodic
protection, are often used with painted surfaces.
Paint systems can be incorporated in boundary element models by defining an
electrical current density of zero for a well painted surface. Specific experimentally
obtained painted surface polarization data can also be used if a more exact solution
is required. Areas of paint damage can be modeled by assigning metal polarization
response to specified elements. Mesh size can become an issue when modeling a struc-
ture which has small areas of paint damage. Extremely small damage on a relatively
large structure may require combining local and global models for accurate represen-
tation. Location of paint damage can be as crucial as the extent of paint damage.
Identification of paint damaged areas should be based on observations from full scale
structures whenever possible. In the case of damage due to erosion, impingement or
other effects which may occur randomly along the structure, statistical approaches
may be required to determine representative damage locations.
The descriptions of protection systems given are brief summaries. The design of
a corrosion protection system is a complex task. Detailed information on the three
systems described as well as other preventative measures can be found in the literature.
62 Industrial Applications of the BEM

5 Example computer modeling of corrosion systems

The ability to use computational modeling techniques to evaluate electrochemical

corrosion has been established. Finite element and boundary element methods have
been used to model geometrically complex structures ranging from tube-to-tube-sheet
connections in steam generators 12 to offshore piping platforms 13. Reviews by Nazami~
Porter and Mufti5, Munn6 and DiGiorgi and Kaznoff14 show the versatility of com-
puter modeling as applied to corroding systems.
Computer modeling techniques can be used to determine the electrical current
density-electrical potential values for the entire surface area of a structure. Boundary
conditions of the first, second and third type are combined to represent a particu-
lar structure, the surrounding medium, and all associated electrical sources or sinks.
Polarization response is used by the computer solution algorithm to correlate electri-
cal current levels and electrical potential values for the exposed areas of metal. The
polarization response is the material constitutive response required for the numerical
solution and can be thought of as equivalent to the stress-strain response required for
structural evaluations. As with any constitutive response, inaccuracies or inappropri-
ate values will result in variation between computational and observed results.
Any computer method which can be used to solve the governing equations can be
used to predict electrochemical behavior. Solution accuracy may be identical for dif-
ferent analysis methods, i.e. finite element and boundary element, however the ease of
computer model formulation may be very different15 . It may be simpler to model the
physical characteristics of a specific structure experiencing electrochemical corrosion
using a particular modeling technique. In the case of a structure submerged or buried
in an relatively large volume of electrolyte, boundary elements techniques are a more
natural choice. The volume of interest which must be modeled to determine elec-
trochemical behavior is that of the electrolyte surrounding the structure. The metal
surface of interest is that at the solid-electrolyte interface. Modeling using finite dif-
ference techniques is restrictive with respect to the degree of geometric complexity
which can be achieved. Often the structure to be evaluated is not geometrically simple
and either boundary element or finite element techniques are required to accurately
model the structure. Modeling using finite element techniques becomes tedious and
complicated because geometric complex three-dimensional models must be generated.
Boundary element techniques require only modeling of the structure-electrolyte in-
terface instead of the entire electrolyte volume. Therefore significant reductions in
model complexities are possible when boundary element techniques are used rather
than finite element techniques. In addition, the model created is a two-dimensional
shell of the structure rather than a full three-dimensional representation of the volume
surrounding the structure. The outer limits of the electrolyte are modeled as the outer
edges of a box or other simple structure completing the boundary element model. In
the case of a structure in an infinite or semi-infinite volume of electrolyte the result-
ing two-dimensional shell type boundary element model is typically a simpler model
than would be required by for finite element evaluation of the structure. Modeling
concessions necessary for the creation of any computer model, which may result in
loss of accuracy, will not be as severe for the simpler boundary element model.
In this section, two example are presented which demonstrate the application of
 Corrosion basics and computer modeling 63

boundary element techniques to the evaluation of structures in infinite electrolyte

domains. The infinite electrolyte domain is representative of structures set afloat in
deep water or buried in thick soil strata. In both of these cases, the extent of the
electrolyte surrounding the structure is large in comparison to the size of the struc-
ture. In the first example, a pipe of two different materials is submerged in a tank of
seawater. The galvanically generated electrical current is determined for a series of
different connection geometries. The parametric study performed demonstrates the
ability to use computer modeling as a design tool to determine the configuration with
the lowest electrical current and therefore the lowest rate of corrosion. In the second
analysis, the impressed current cathodic protection system of a ship is modeled and
the electrical current required to maintain protection against corrosion is determined.
The second analysis addresses such issues as multiple material properties, complex
geometries and comparison with experimental results.

5.1 Piping system evaluation

One of the advantages of computer modeling is the capability of comparing system

performance. This capability allows the designer to consider the effects of changes in
system design on overall system performance without the cost associated with per-
forming an experimental evaluation of each configuration. The boundary element
technique example involves the determination of the most beneficial connection of
two galvanically incompatible materials in a piping system which is submerged in sea-
water. A parametric approach is used to determine the most desirable configuration.
The ideal piping system would be fabricated from one material and if two ma-
terials were joined they would be electrically compatible minimizing the resulting
galvanic corrosion. Piping systems, however, often consist of several materials which
are electrically incompatible resulting in galvanic corrosion damage to the piping sys-
tem. Piping systems commonly include a variety of connections such as piping of the
same diameter but different materials as well as pieces of piping of different diameters
which may or may not be of the same material.
The boundary element technique example problems is the evaluation of possi-
ble connections between two pieces of pipping of different materials. The two metals
chosen for study are Inconnel and Copper–Nickel (CuNi) alloy. These particular met-
als were chosen for the demonstration problem because the difference in open circuit
potentials and polarization response will result in galvanic corrosion. It is the intent
of the computer modeling and analysis to determine how galvanic corrosion can be
minimized for these two materials and not whether significant galvanic corrosion oc-
curs. A commercial boundary element code was used in all evaluations. There were
no external applied electrical current or voltages. The only driving force used in the
analysis was the difference in material polarization responses for the Inconnel and
CuNi alloy. The analysis used standard program features and no special computer
coding to analyze the piping.
A straight pipe consisting of equal lengths of the 10.2 cm diameter Inconnel and
CuNi alloy was evaluated. The total length of the pipe is 934.7 cm. The pipe is sub-
merged in the center of a 914.4 cm diameter tank filled with seawater. A boundary
element model of 260 elements and 463 nodes was created of the exterior surface of
64 Industrial Applications of the BEM

pipe and the surrounding seawater. Quadratic elements were used to model the pip-
ing sections, linear elements were used to model the tank boundary. Small elements
were used in the region near the connection between the two materials in order to ac-
curately capture the electrical current density and electrical potential gradients near
the material interface. Axisymmetric modeling was used so that the model, shown
in Fig. 4, represents the outline of a 1° slice of the circular piping and surrounding
seawater. The volume of seawater surrounding the pipe is large but finite. Nonlinear
polarization response curvesll were used to characterize the material behavior. The
seawater was modeled as a homogeneous electrolyte with a constant conductivity of
0.05 Ohms-l-cm-l.
The electrical current density does not vary significantly along the anodic section
of piping with the maximum electrical current density of 4.8 x 10–6 to Amperes/cmz
occurring at the material interface. The approximately constant electric current den-
sity will result in corrosion which will be uniform along the length of the CuNi section
of piping. The amount of corrosion at any location can be approximated by that
caused by the maximum electrical current density. The corrosion rate can be de-
termined using Faraday’s law and the appropriate material properties, as shown in
Eq. (7). Based on the maximum electrical current density, the corrosion rate for the
submerged pipe is 0.11 mm/yr.
The second pipe to be evaluated is fabricated from equal length sections of 30.4cm
diameter Inconnel pipe and 10.2 cm diameter CuNi pipe. The total length is 934.7
cm. Four different piping connections were considered; a step connection, and a 121.9
cm long taper of Inconnel. The piping was submerged in a 914.4 cm diameter tank
filled with seawater. The pipe and the tank were modeled using an axisymmetric
boundary element model of 445 linear elements and 450 nodes was created of the step
connection geometry. Element size was larger than that of the previous two material
straight pipe model and the increase in element size was based on the results of the
previous analysis.
In all cases, step and taper connections, the maximum electrical current density
occurs in the CuNi alloy, the anode, at the material interface. The maximum electrical
current densities for the different connection geometries are:
— step connection – 9.7 x 10–6 Amperes/cm2
— Inconnel taper -12.0 x 10-6 Amperes/cmz

- CuNi taper – 8.5 x 10-6 Amperes/cm2

It is apparent from the maximum electrical current density that the piping connection
which would result in the least amount of galvanic corrosion is the taper connection
with the taper fabricated from CuNi alloy. The oxidation rate of the CuNi alloy must
balance the reduction rate of the constant surface area of Inconnel. Therefore, the
lowest corrosion rate will occur for the connection which has the smallest cathode to
anode area ratio.
As demonstrated in this example, it is possible to evaluate the performance of
different piping system configurations once materials have been selected. While not
demonstrated in the study presented, it is possible to generate corrosion rate infor-
mation for different material combinations for similar piping system geometries. It is
 Corrosion basics and computer modeling 65

also a relatively simple matter to apply the parametric study approach presented to
more complex structures. While it may not be possible to prevent all corrosion related
failures due to the complex interaction of structure and environment, it is possible
to predict failure in many cases when corrosion evaluation is included in the design
process. The simple studies performed show the versatility of boundary elements and
computer modeling as a design tool.

5.2 Impressed current cathodic protection system analysis

Evaluations of an existing impressed current cathodic protection (ICCP ) systems on

U.S. Navy CG class surface ship were performed using commercial boundary element
code. Electric current requirements for a six anode single zone and a six anode two
zone system are evaluated for static (ship at rest) and dynamic (ship underway)
conditions16 The example presented in this section is the evaluation of the six anode
single zone system for static conditions. A zone is an independently controlled power
supply used to provide the electrical current to the impressed current anodes. Anodes
within a control zone have identical voltages and electrical current values. Therefore,
the six anodes of the system studied have identical voltage values assigned. The
magnitude of electrical current for each anode is used for a system solution check.
The magnitude of electrical current is calculated by the boundary element program
and should be equal for all anodes.
The computer modeling results are compared with experimental data obtained
through physical scale model testing. Physical scale model testing involves the scaling
of geometric dimensions of the structure and scaling of the electrolyte conductivity 17.
Physical scale modeling experimental results have been successfully compared with
data obtained from tests performed on full size ships17. The ability to use scale
model test results, which can be related to full size ship data, increases confidences in
the reliability of computational results. The physical scale model experiments were
performed using a scaling factor of 1/96 for the models and seawater conductivity.
The ship geometry and ICCP system examined is shown in schematic form in
Fig. 6. The boundary element model consists of half of the hull geometry taking
advantage of port and starboard symmetry to reduce the mesh size. Only the portion
of the ship hull bellow the water line is required to be modeled. The boundary element
model of the CG class surface ship consists of 573 quadrilateral constant value elements
and 700 nodes (Fig. 6).
The boundary element model of the CG class ship geometry developed has the
following features:

- detailed modeling of the complex curvature of the hull surface

— inclusion of rudder as geometrically distinct but electrically connected compo-

nent

inclusion of propellor assembly as geometrically distinct but electrically con-

nected component

use of nonlinear polarization response curves to represent material behavior

66 Industrial Applications of the BEM

— use of three separate materials (steel, nickel–aluminum–bronze (n–a–b) and

paint) to model separate regions of the ship.

The impressed current anodes have finite areas and fixed locations and are explicitly
included in the boundary element model through use of constant voltage boundary
conditions. The total electrical current into the system through the impressed current
anodes and the total electrical current leaving the system through the exposed metal
sections are compared as a check on the accuracy of the calculated solution.
Three different materials are used to model the hull, propellor and rudder in the
model. The hull is assumed to be undamaged paint and therefore no electrical current
is allowed to be transmitted to the hull surface. The propellor and rudder are modeled
as bare nickel-aluminum-bronze (n-a-b) and steel, respectively. Nonlinear representa-
tions of polarization response were used for steel and n-a-b. Seawater is defined as a
homogeneous elect rolyte with a constant conductivity of 0.05 Ohms-l cm-l.
The material configuration of the physical scale model test specimens and the
surface ship as represented by the computer models are different. There are four differ-
ences between physical scale model test ship and computer model material definitions.
These differences must be taken into consideration when comparing global and local
results. The first two differences are a result of mesh refinement limits resulting from
time and financial limitations. The scale model test ship contains the bilge keel. The
boundary element does not. The bilge keel can be idealized as a narrow ridge which
is perpendicular to the hull as shown in Fig. 5. The bilge keel acts as a deflector and
shields a portion of the hull from electrical current flow and the associated electrical
potential of the seawater at the hull surface. The boundary element does not contain
the bilge keel because it was felt at the time of original modeling that the bilge keel
does not act as a significant geometric feature. Subsequent work indicates that this
may not have been a valid assumption. The physical scale model test ship also con-
tains docking blocks which represent 1910 of the total surface area. The docking blocks
are bare steel and are located along the port-starboard centerline. The docking blocks
were not included in the computer model because of the mesh size limits resulting
from time and financial limitations. Docking blocks are relatively small geometric
features and it was felt that they were smaller than the level of detail required. The
third and four differences in the two models deal directly with the representation of
the rudder and propellor. The rudder in the physical scale model test ship is painted.
The rudder in the boundary element model was bare steel. It was the intent of the
original computer model to represent extreme damage to the rudder. Finally in the
boundary element model the propellor is modeled as a solid disk as opposed to the
detailed propellor of the physical scale model test ship. These variations in material
assignments for different components will effect the tot al required electrical current
because of differences in exposed metal areas.
It is often unfortunately the case that experimental and computer models are not
identical. The reasons for differences in the two models can vary from miscommuni-
cation to inability to create as refined a computer model as necessary. The analyst
must recognize the areas of difference and be able to identify results which can be
compared. The ideal situation occurs when there is an one-to-one correspondence
between experimental specimens and computer models. In this case, the value which
 Corrosion basics and computer modeling 67

can be meaningfully compared is the total electrical current absorbed by the propel-
lor.
ICCP systems are controlled by defining a target electrical potential associated
with a specified reference cell. Corrosion will not occur when steel is catholically po-
larized to an electrical potential of -0.85 volts measured with a Silver-Silver Chloride
(Ag/AgCl) electrode. Physical scale model experimental results used for comparison
are for reference cell readings of -0.85 Ag/AgCl. In the boundary element analysis,
a reference cell potential range of -0.83 to -0.87 volts Ag/AgCl is used to define the
target condition. In order to bound all possible solutions, upper and lower bound solu-
tions are defined based on electrical current requirements. The upper bound solution
is defined as that which requires more electrical current and is therefore associated
with a reference cell reading of -0.87 volts Ag/AgCl. One design problem associated
with ICCP systems is the ability to locate a reference cell which is truly representative
of the hull surface condition. It is also possible to determine, by examination of all
hull surface potentials, how well the reference cell reading represents the potential of
the entire hull surface.
Since the boundary element program uses an iterative solution technique it is
possible that the electrical current values associated with impressed current anodes
may not be exactly equal for anodes within a zone. The solution tolerance will effect
the variation in currents determined as part of the boundary element solution. Be-
cause of the iterative nature of the solution procedure a tolerance on electrical current
equivalency is required to determine solution acceptance with regards to the criteria
imposed on each power supply zone. The impressed current anodes belonging to the
same zone must have a total electrical current input within 0.5 Amperes of the other
anodes attached to the same power supply for the solution to be considered valid.
The current and voltages required for a lower bound solution (reference cell at
-0.83 volts Ag/AgCl) for the single zone system at static conditions are:

VA = V’ = VC = –1.3 volts, 1A = IB = Ic = 3.2 to 3.4 A m p e r e s (16)

where the locations of anodes A, 1? and C are as defined in Fig. 5. The range in
electrical current for anodes is due to the iterative nature of the boundary element
technique. A more refined mesh and tighter convergence tolerances will eliminate this
variation. The negative voltage value indicates voltage supplied to the system for the
particular boundary element formulation used in the analysis.
The upper bound solution (reference cell at -0.87 volts Ag/AgCl ) requirements
are:
VA = VB = VC = –1.45 volts, 1A = IB = IC = 3.7 to 4.0 Amperes (17)
Electrical potential profiles for the port-starboard centerline and for the side of the
hull at a depth of 3.0 m. are shown in Fig. 7. Computer and physical scale model
test data show similar trends.
Physical scale model and computer results indicate that electrical potential values
on the rudder and propellor are less than required for corrosion prevention despite
the electrical potential at the reference cell. The range in electrical potential on the
rudder side tip of the propellor is -0.76 to -0.82 volts Ag/AgCl. The physical scale
model measured electrical potential for the same location is -0.72 volts Ag/AgCl.
68 Industrial Applications of the BEM

Despite the good agreement in electrical potential values, both along the two pro-
files chosen and for a specific point, there is a large discrepancy in the total electrical
current calculated from physical scale model tests and the computer modeling solu-
tions. Boundary element solutions result in a range of total current required from
20.2 to 23.8 Amperes. Of this total, 10.0 to 10.6 Amperes is delivered to the propel-
lor. Physical scale test data indicates a total current requirement of 63.9 Amperes
with 43.1 Amperes delivered to the propellor. Possible explanations of the difference
in the amount of electrical current delivered to the propellor are the uncertainty in
the material polarization response curves used in the computational model and the
presence of the bilge keel in the physical scale model. The n-a-b polarization curve
was determined using filtered seawater and small laboratory test specimens 1. The
laboratory data used may not be the most appropriate for the conditions and material
surface features of the n-a-b used in the physical scale experiments. It is possible that
more appropriate polarization curves may result in increased accuracy of computer
results. The bilge keel has been observed to act to deflect the flow of electrical current
and may be partially shielding the reference cell in the experiments.
Based on the analysis performed the boundary element method has been demon-
strated to accurately predict global behavior. Differences between experimental and
computational results can be attributed to variation inherent in the material polariza-
tion response and to geometric characteristics not included in the boundary element
model. Additional mesh refinement, which will allow for inclusion of the docking
block and bilge keel features should also result in an improvement in total electrical
current requirements. It is interesting to note the relative sensitivity of the electrical
potential profile and total electrical current. Electrical potential profiles, which are
a global structure response, are less sensitive to mesh refinement and material char-
acterization issues than total electrical current, which is the integration of the local
current density response.
Once the desired accuracy of computer modeling is obtained and verified by com-
parison with experimental data, the versatility of computer modeling can be exploited
by the analysis. Power supply zones connecting various anodes can be created by
changing the defined voltage or electrical current or current density boundary con-
ditions for anode elements. By simply changing the value of boundary condition
definitions the analyst can explore all possible power zone combinations and power
source limitations. The reference cell location, which may or may not yield electrical
potential values which are indicative of potential levels in regions of interest such as
the propellor, can be changed to any existing node location and need not be pre-
define. Electrical potential profiles and contours which show the global response of
the ship can be used to determine system performance. Anode placement can be
altered by redefining prescribed boundary conditions to different elements. Different
operating conditions, defined by temperature, salinity or ship speed, can be evaluated
by changing the resistivity of seawater and the material polarization response. The
list of possible combinations is, for all practical purposes, endless.
 Corrosion basics and computer modeling 69

5.3 S u m m a r y

The two examples presented in this section demonstrate the versatility of the bound-
ary element method. The ability to modify existing systems, whether by changes in
materials or geometry, is a powerful aspect of computer analysis. The examination of
beneficial or adverse effects of design variations, intentional or as a result of fabrication
difficulties, are another positive feature of computer analysis. While understanding of
the mathematics of the computer programming and of the physical phenomena is nec-
essary for the successful completion of a computer analysis, this should not deter one
from the use of this powerful tool. Commercial boundary element computer codes are
readily available or it is possible to create special purpose codes within a reasonable
amount of time based on the existing technology. In summary, the corrosion charac-
teristics of a system may be determined using existing computer modeling techniques.

References and Suggested Reading

References

1. Economic Effects of Metallic Corrosion in the United States, National Bureau of Stan-
dards, NBS Special Publication 511-1, U. S. Dept. of Commerce, May 1978.

2. Fontana, M. G., Corrosion Engineering, McGraw Hill, New York, 1986, p. 1.

3. Tomashaw, N. D., Theory of Corrosion and Protection of Metals, The Science of

Corrosion, The Macmillan Company, New York, 1966, p. 5.

4. Peterson, M. H., Physical and Mathematical Modeling of Current Distribution and

Cathodic Protection, 8th Inter-naval Corrosion Conference, Proceedings, Plymouth,
England, 1988, p. 1-9.

5 Zamani, N. G., Porter, J. F. and Mufti, A. A., A Survey of Computational Efforts

in the Field of Corrosion Engineering, International Journal of Numerical Methods in
Engineering, Vol. 23, 1986, p. 1295-1311.

6. Munn, R. S., A Review of the Development of Computational Corrosion Analysis

for Spatial Corrosion Modeling Through Its Maturity in the Mid-1980 s- , Computer
Modeling in Corrosion, ASTM STP 1154, American Society for Testing and Materials,
Philadelphia, 1991, p. 215-228.

7. Jones, D. A., Principles and Prevention of Corrosion, Maximilian Publishing Co.,

New York, 1992, p. 118.

8. Jones, D. A., Principles and Prevention of Corrosion, Maximilian Publishing Co.,

New York, 1992, p. 92.

9. Corrosion Basics, An Introduction, National Association of Corrosion Engineers, Ho-

uston, 1984, p. 151.
70 Industrial Applications of the BEM

10. Jones, D. A., Principles and Prevention of Corrosion, Maximilian Publishing Co.,
New York, 1992, p. 123.

11. Hack, H. P. and Scully, J. R., Galvanic Corrosion Prediction Using Long and Short-
Term Polarization Curves, Corrosion, Vol. 43, No. 3, 1986, p. 79-90

12. Scully, J. R. and Hack, H. P., Prediction of Tube-Tubesheet Galvanic Corrosion Using
Finite Element and Wagner Number Analysis, Galvanic Corrosion, ASTM STP 978,
Philadelphia, PA, 1988, p. 136-157.

13. Strommen, R. D., Computer Modeling of Offshore Cathodic Protection Systems:

Methods and Experience, Computer Modeling in Corrosion, ASTM STP 1154, Amer-
ican Society of Testing and Materials, Philadelphia, 1992, p. 229-247.

14. DeGiorgi, V. G. and Kaznoff, A. I., A Review of Computational Simulation Tech-

niques, Corrosion Forms and Control for Infrastructure, ASTM STP 1137, American
Society for Testing and Materials, Philadelphia, 1992, pp. 77-89.

15. Fu, J. W., Numerical Modeling for Galvanic Corrosion and Cathodic Protection Anal-
ysis, Computers in Corrosion Control, Proceedings of the Corrosion/86 Symposium
on Computers in Corrosion Control, National Association of Corrosion Engineers,
Houston, TX, U. S. A., p. 161-175.

16. DeGiorgi, V. G., Lucas, K. E., Thomas, E. D. and Shimko, M. J., Boundary Element
Evaluation of ICCP Systems Under Simulated Service Conditions, Boundary Element
Technology VII, Computational Mechanics Publications, Southampton, 1992, p. 405-
422.

17. Thomas, E. D. and Parks, A. R., Physical Scale Modeling of Impressed Current Ca-
thodic Protection Systems, Corrosion 89, Paper 274, National Association of Corrosion
Engineers, Houston, TX, 1989.

18. Thomas, E. D., Lucas, K. E., and Parks, A. R., Verification of Physical Scale Model-
ing with Shipboard Trials, Corrosion 90, Paper 370, National Association of Corrosion
Engineers, Houston, TX, 1990.

Suggested Reading

Theoretical

● Fontana, M. G., Corrosion Engineering, McGraw Hill, New York, 1986

● Jones, D. A., Principles and Prevention of Corrosion, Maximilian Publishing Co.,

New York, 1992

● Tomashaw, N. D., Theory of Corrosion and Protection of Metals, The Science of

Corrosion, The Macmillan Company, New York, 1966
72 Industrial Applications of the BEM

Galvanic Series

Metals Exposed to seawater

Noble (Cathodic)
Platinum

Gold
Titanium

Silver

Type 316, 317 Stainless Steels (Passive, with Oxide Film)

Nickel
Nickel-Aluminum-Bronze
7030 Copper Nickel
Lead

Tin Bronzes
Silicone Bronzes

Manganese Bronze
Copper
Tin

Naval Brass, Red Brass, Yellow Brass

Type 316, 317 Stainless Steel (Active, without Oxide Film)
Low Alloy Steel

Mild Steel, Cast Iron

Aluminium Alloys
Zinc

Magnesium

Active (Anodic - Corroding Surface)

Table 1
 (+) Activation Polarization

Tc P

(-)

Log Current

Limiting
(+) Concentration Current
Polarization

~c

(-)

Log Current

Figure 1
74 Industrial Applications of the BEM

Measured
Polarization
(+) ,,,,,,/,,~,//////0.0.,/.,.,,,

llc

(-)
.
IL
Log Current

Figure 2
Corrosion basics and computer modeling 75

t
Transpassive

Passive

—! ,
1
Active
76 Industrial Applications of the BEM

Physical Geometry

CuNi

Seawater

Pipe Radius = 5.1 cm

Tank Radius = 457.2 cm
Total Pipe Length = 934.7 cm

Figure 4
 Corrosion basics and computer modeling 77

/
%“ Bilge Keel
Rudder
Propellor
Source Anodes - A, B, C
Reference Cell - R

Figure 5
78 Industrial Applications of the BEM

Figure 6
 -1.5

‘1.25

-1
l-l
‘AtiAK!o.75

-0.5

-0.25

Frame

Depth = 3.0 m.
-1.5

-1.25

-1
n ~
(Ag/AgCl) ●
-0.75

-0.5
-0.25