97512 Corrosion Monitoring and Control in Refinery Process Units (51300-97512-Sg)

7/21/2019 97512 Corrosion Monitoring and Control in Refinery Process Units (51300-97512-Sg)

http://slidepdf.com/reader/full/97512-corrosion-monitoring-and-control-in-refinery-process-units-51300-97512-sg 1/25

Paper No.

5

CORROSION MONITORING AND CONTROL

IN REFINERY PROCESS UNITS

Alec Groysman

Oil Refineries Ltd.

P.O. Box 4

31000 Haifa, Israel

Avihu Hiram

HIRAM - Process Control Eng. Ltd

22 Martin Buber St

34861 Haifa, Israel

eMail:hiram@netvision. net.il

ABSTRACT



INTRODUCTION

Crude oil contains different species which while beii processed reacted or converted

into substances corrosive to alloys.

Water is condensed during the production processes and gases like hydrogen chloride

(I-ICI) and hydrogen sulfide (H,S) are dissolved in this water, various salts (chlorides,

sulfides, hydrosulfides) and oxide-hydroxide deposits formed on the metal surhces result in

severe corrosion in the overhead system of the crude distiition unit. Air coolers, heat

exchangers, condemers and pipes made of carbon steel suffkr km general corrosiorq

pitting corrosio~ and under deposit corrosion.

Prevention of corrosion damage and possible detrimental eflkcts on the environment

depend on knowledge of the corrosion situation at the units, that is “corrosion behavior” of

metallic equipment, especially their corrosion rates and corrosion forms.

Various types of corrosion monitoring methods were recommended fix follow-up in

the overhead system of dkdlation units [1,2]. We did not tind literature comparing the

various types of monitoring results and how ef%ctive it is in evaluating real time corrosion.

SEM & EDS method is widely used for the identitlcation of corrosion products [3],

but we did not find in the literature, how this method is particularly used fir the chemical

idedfkation of corrosion products and morphology of coupons’ surfhces f.iom the

overhead system.

Given this

apparent

lack of well published basis for using various corrosion

monitoring techniques, we attempted to use our practical experience to define ways how to

utilize these tools to monitor Real Time Corrosion in the Equipment.



EXPERIMENTAL PROCEDURE

A. MASS LOSS ML) METHOD

Retractable corrosion probes with coupons were used, The mass loss of coupons

made of a carbon steel (the same material the corrosion of which we want to monitor) strip

specimen and their corrosion rates were measured every 40-60 days. The exposure period

depended on the corrosion, In some cases, where the corrosion rate was too low to be

measur@ we had to extend this period up to 190 days.

After the retraction of the coupons, hydrochloric acid (5%) with organic corrosion

inhibitor was used for cleaning coupons’ surthces from corrosion products and W.

Analytical weighing was used for definition of the mass loss of the specimen.

B. ELECTRICAL RESISTANCE ER) METHOD

Retractable ER-probes were used fix measuring changes of electricrd resistance of

sensor made of carbon steel wire or cylinder and thus corrosion rate. The ER-probes were

located m close proximity to the corrosion coupons, close to the wall of heat exchangers or

pipes as possible. The measuring of changes in the electrical resistance of the sensor was

done by portable device or continually monitored by an on-line monitoring (acquisition)

system.

The corrosion rates were

calculated

from the slope of line “Dial reading of ER-probe

versus Measuring time”.

The on-line monitoring included also data collecting systeq calculations, performing

and utilization of received data.

C. CHEMICAL ANALYSIS OF ACCUMULATED CONDENSED SOUR WATER



‘The film and deposit thickness on the specimen was measured by means of special

device for measmhg non-conductive or non-magnetic ilhns on the iron. Some expertise and

experkme needed in order to make out the important tits which should be considered in

the anti-corrosion efforts.

E. THE IDENTIFICATION OF CORROSION PRODUCTS ON THE COUPONS

The chemical composition of corrosion products, films and other deposits formed on

coupons’ surhces as well as morphology of these surfhces after removing of deposits was

M by IIBXUISf SEM & EDS. The device JSM-5400 of JEOL was used fix these

purposes.

RESULTS AND THEIR DISCUSSION

The arrangement of corrosion probes in the overhead systems of a typical crude

distillation units is presented m Figure 1. The general scheme of corrosion monitoring logic,

the relationship betsveen its various methods and its posslMlities are descrii in F~e 2.

I. CHEMICAL ANALYSIS OF ACCUMULATED SOUR WATER

‘h chemical analysis of accumulated sour water formed after condensation of

overhead gases from the atmospheric dihlhtion colurnm+has been used for many years to

make out the details about corrosion “behavior” in the overhead systems. The chemical

analysis of the corrodents (p~ chlorides, sulfides, suliktes) and the corrosion product

(dissolved iron mainly) give relatively fhst and reliable “picture” of both sides of the

corrosion system. These data fimilitate improving neutralization and inlrdition treatment in

the overhead system and improving the desalting process of the crude.

The chemical data of accumulated sour water do not reflect exactly the real corrosion



High changes of pH values as well as content of chlorides and iron in sour water

indicate an unstable conditions in processing and as a result bad corrosion “behavior”. In

spite of this, there is no iniiormation about the real corrosion rates and the actual corrosion

forms. There is no way to forecast, km the chemical analysis of the sour water only, the

tube service life and shut-down periods.

The chemical method gives qualitative rather than quantitative information of the

corrosion situAon. Therefore the metal loss coupons, visual examinatio~ identification of

corrosion products and iilrns and ER - probes were utilized to complete the “picture” of the

corrosion in the overhead system of the crude distillation units.

II. MASS LOSS ML) METHOD

This method is descriid in literature [4,5]. There is no universal method for

corrosion follow-up. Every method has advantages and d~vantages, limits and benefits.

Mass loss method enables memwing the average corrosion rate over a period. Visual

-on

enables detining the presence of deposits and pits. The salt deposits on the

coupon aurt%cesshows the necessity of increase washing m the overhead system.

Chemical

identiktion of films and deposits firmed on the surf%cespecimen enables

ascertaining corrosion mechanisms, that is the causes of corrosion process and thus

Cbinkh@ or preventing its spreading.

The results of mass loss method corrosion follow-up that have been collected for the

last two years are presented in F- 3.

The strip coupons show the corrosion situation only in the places of its mounting. The

corrosion situation may dif.lkrin the other places of the system. Therefore the more coupons

we insert in the overhead system the better “picture” we are going to get of the real



data. The thickness of tube bundles m many cases is 2.336 mm. The planned Me of air

coolers, condensers and heat exchangers m refineries is 15 years [6]. Thus the accepted or

permissible corrosion rate for this equipment is:

0.7 * 2.336 Lnrn/15y = 0.11 MM/y(5 my),

where 0.7 - wear and tear coefficient, that is the

maximum percent (70??) that

thickness of tubes are allowed to be diminished up to its replacement. It means that

corrosion rate up to 0.11 nudy is acceptable.

This value cotildes with the one NACE recommendation as an accepted low

corrosion rate for the equipment in the oil industry [4].

SEM EDS analysis of corrosion products and films formed on coupons

In addition to corrosion rate values, coupons enable defining corrosion form and

mechanism based on surfhce state (morphology) and chemical identification of films and

corrosion products formed.

General corrosion with uniform tilms, and pitting corrosion under non-unifbrm films

and deposits have been found.

The ardysis of corrosion products made by means of SEM & EDS, showed the

presence of iro~ suli%r, oxyg~ s.msll quantity (traces) of chloride and sometimes even

bromide (Figures 4,5).

The obtained

data

point

out

that

iron suliides (probably FeS and some other forms

FeJ+.) covering the carbon steel surfiwe are responsible for the formation of protective iihn,



were Rmned under non-unifbrm films and deposits. The iron sulfide iilm is cathodic to iron.

If this film were perfect, the galvanic pair does not work, therefore corrosion is negligible. If

this tl.hnwere non-unifiormand impmfkct, the corrosion agents ingress in impmtkctions and

severe pittii corrosion occurred under ilms.

The presence of annnonia or amines and chlorides m the overhead system results in

the formation of ammonium and amine chloride deposits which may hydrolyze with

formation of hydrochloric acid under deposits.

The pH values decrease to 2-3 and results in corrosion under deposits and m its turn

pitting corrosion. The corrosion rate under deposits can be as nmch as 100 times as the

general corrosion rate. In order to prevent the formation of these deposits, ammonia usage

is mininimd and generous water wash is applied. The deposits consisted of chloride salts on

the coupon surfhces is an indicator of wash quality.

In cases where the deposits and sediments were consisted of iron sulfides water wash

could not be effbctive, because the sulfide salts are not water soluble.

The mass

loss

method has some limitations.

Corrosion coupons examination are unable to ditlkrentiate between long term steady

corrosion rates and corrosion which has occurred rapidly over a short period of time, due to

upsets or other events.

This method gives the overall average corrosion rate during some, usually long

period. If corrosion rate was low (thousandths mm/y), the exposure time should be

prolonged up to as long as halfa year and it is impossible to retrieve any data during all this

Verylong period.



Curve fitting by regression analysis was carried out for the data received from this

device.

ER-method gives results considerably quicker, aud less effimts are needed m

comparison to the mass loss method.

Properly selected ER-Probes enable data to be read close enough to give instant

information about events m the “ER-probe litk”. The various results may be received for the

M%rent periods (Figure 7). Phenomena like sudden reversing the corrosion process might

be indicated R values go to the opposite direction - decreasing @ii 8). Finding

reasonable explanation for this phenomena is always problematic. One of the explanations

for such fluctuations is that protecting tlrns on the sensor surilice are formed and destroyed

periodically. Another possibility is the periodic formation of films with various electrical

conductivity. If the memring

period taken is long (the same period as fbr mass loss

method) then identical results are received.

As a rule the corrosion rates are lower when measuring periods are longer (see Fig.

8). The similar results may be received by means of the mass loss method. When the

exposure period of coupons is increased the average corrosion rate is diminkhed. Generally

this relationship occurs when dependence of specimen mass loss versus time is descrii by

a curve with “saturation” (Fii 9). This phenomenon often occurs when carbon steel is

under neutral or near neutral solutions. Comparison of the results received by means of two

methods, mass loss and E~ during long-term periods shows their fidl codhnity (Figure

lo).

B. On-line corrosion monitoring.

A number of ER-probes with 4-20 ma output as a fimction of R changes were

installed m the overhead of two crude distillation units (see F- 1). The installation of



The analog signals (4-20 ma) are sent from the probe are proportional to the un-

corroded matter left and are calibrated to give an indication of total corrosion of the steel

sensors in microns.

The outputs of these probes are monitored via the existing DCS

(Honeywell/TDC3000) and are also sent to the Plant Mormation System (PI) where they

are analyzed. Figure 11 shows these data as collected and stored within the PI System. The

PI System enables easy access to the history data base with every simple PC which is

connected to the local net or by dialing-in tiom remote PCs. Bringing this intbrmation onto

the PC makes possible the manipulating with all kind of analysis tools not only from within

the PI System but from outside it on other systems which might be available on any PC in

the network (e.g. commercial available spreadsheets and various Stattilcal software).

Performing the cakulations.

Based on the field imtrume nt readings corrosion rates

are calculated by means of WRmmtiation of received data regarding given period. In order

to make mean@@d calculations it is essential to fist tilter the noise from the data. A

number of approaches were practiced.

The simplest way is to take all the readings that were collected within the Mormation

System and find the regression line which best fits the data (Figure 12).

Another technique which is more suitable fbr on-line calculations is Mtering by

averaging over periods of time and making the corrosion calculations between separated

averages. There are two parameters to adjust in the application of this algorithm the length

of the period to be averaged to give representing values and the time to separate between

these tsvo values (Figure 13).

The calculated corrosion rates are depended on measured period and these data are



The tube bundles in the overhead atmospheric unit are under reduction conditions

~zs atmoS@e@ during nod

operation. men the equipment is opened for inspection

they are exposed to normal atmosphere (with oxygen), and oxidation processes immediately

take place. The black deposits of iron sultide inside the air cooler becomes reddish iron

hydroxides within several hours after opening of the cooler. Therefore it is important to

have the inspection of the tube bundle done as soon as it is opened to the atmosphere.

Visual exarnina

tion had been done immediately afler opening the heat exchangers. The tubes

were found to be covered with black deposits of various thickness. The chemical analysis

showed the deposits consisted of iron sulfides and iron oxides. These deposits were

removed bywater jet, and tube surt%cewas then visually inspected.

The general condition of the tubes’ sties was good, but shallow pits of various size

were discovered under the thick deposits. The pits on the tubes were very mild and of no

dangerous for the on going service of the heat exchangers. These pits required no special

attention and no speci6c action to be taken

The actual equipment inspection findiis matched the available on line monitoring

data.

The ability to react i%stto the signals given by the on line monitoring system enabled

the anti-corrosion program to be adjusted in real time and helped to maintain the equipment

under good service conditions.

CONCLUSIONS

1. The experience of corrosion monitoring in the overhead system of the crude distillation

units in the reiinery has been described.



when deposits and non-tiorm Wns of more than 80-100 microns thickness were

formed.

7. The actual equipment inspection iindings showed general good condition of tubes

bundles and thus matched the available on linemonitoring data,

8. The availab@ of on-line corrosion monitoring system enabled better look after anti-

corrosion treatment and made the overall program much more successfid.

REFERENCES

1. Russel C. Strong, Raymond A. Stephenso~ Ara J. Bagdaaar@ James E. Feather,

Robert B. Hti “Crude Unit Corrosion and Corrosion Control”, CORROSION/96,

paper no. 615, @otiOQ TXNACE Internatio~ 1996).

2. Milton P. Ramo% Luiz A. Corre%

“On-Line Corrosion Control in Refinery

Overhead System”,

CORROSION/95, paper no. 335, Houstonj TXNACE

Internatiod 1995).

3. Thomas Mebrahtuj K.J. Del Rossij “SEM and XPS Characterization of the Carbon

Steel SurfSee Paasivation Film in Anhydrous Hydrogen Fluoride Media”,

CORROSION/95, paper no. 341, Housto~ lXNACE Internatio@ 1995).

4. RP 0775-87 NACE. “Preparation and Installation of Corrosion Coupons and

Interpretation of Test Data in Oiltleld Operations”.

5. ASTM G4-84. “Standard Method for Conducting Corrosion Coupon Tests in Plant

Equipment”.



NH3 N C.I.

111

A

l.66 p

116.

T

N -

t+leutralizi~_q Amine

CJ. - Corrosion Inhibitor ~_401

L

0.]63 ~

E-4(I1

ATM

G13LUMN

U. 1336

1 ?“ )1 b’+)

42.

T

0.;26

0.869 0.;23 0.;30

-1

c-419/l

c-419{2

0. 0 0

4

I

mm~ear

LAST CTIEEK

A 0.063



B

isual

Inspection

of

equipment

Deposita,

Pits

El

oupons

Mass loss

Visual

examination

Corrosion Rate,

Deposits, Pits

rognosis of

equipment,

Tube Service

life,Shut-down

periods

K1

R - robe

Manual

On Line

Corrosion

rat

n

hemical

alysisof

Condensed

SourWater

Corroderd%

IsQ

W 21



c

o

.-

IJl

0

L

b

c1

0.28

0.26

0.24

0.22

0.2

0.18

0.16

0.14

0.12

0.1

0.08

0.06

0.04

0.02

0

N ~ C.I.

1’4+

‘+

4

~

.——

5 MPY

.03.94

m Before

oir .ooIer

2 ~

After condenser 3

Figure 3.

Corrosion rate vs time.



Fe

c

0

u

n

t

s

123456

1200

1100

1000

900

Fe

)~. ,

“

Zn

78 10 11 12 13 14 15 16 17 18 19 20

E%ergy (keV)



160

150

1401

1301

1201

11OI

c 100

0

u

90

: ~o

s

70

60

50C

400

s



520

510

500

490

480

470

460

450

440

430

420

410

400

390

380

370

/

/

K =(aR/~t ) * 0 365

10 MPY

7.3 MPY

u

[

o 10

20

Figure 6.

Dial reading

30 40

50

30,12.93 - 08.03.94

ER-probe vs time.

60

70 Days (t)



0

c

.-

-0

0

1

400

390

380

370

360

350

340

330

320

310

AC:

0,05 r In/v

/

c

o

10 20 30 40 50 60 70 80

90 100 I1O 120 Days

02.09 - 30.12,93

Figure 7.

Dial reading ER-probe vs time.



140

135

130

125

120

115

I

110

105

100

95

90

85 ‘

D

I

10.12-30,12.92

0,10 mm/y

10,12-10,01,93

0,06 mlI1/y

I 10.12 -19.01,9310 .02 mm/y I

I I

—

—

o

10

20

30

40 Days

10,12.92 - 19.01,93

Figure .

Dial reading ER-probe vs time.



Mass

Loss

Corrosion

Rate



0.4

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0

[

1

{p

\\

\\

T

I

1

I

I

1

I

I

I

I

09.08-22.08.9~ 0.12-19.01 .9~4.04-02.06.9 2. O9-O2.l 1

4 12.93–09.03. 4

01.1 1–26.1 1.9227.01 –21.02.9317.06-02 .09.9 2.1 1–30.12.9 9.03–05.05.94

Mass Loss, mm/y

@ ER–probe, mm/y

Figure {0. Corrosion rate (two methods) vs time,



1

$

II

t1(

1

.

.

A

.

”

”

e

”

.

-

A

L

L

1

1

11

1

I

I

1

f

“

1

1

1

1I

$

1

t

1

“

”

.

-

+

.

+

.

I

1

1

1

1

)

$

1

s

k

1

t

I

1

1

.

L

”

L

”

m

-

Ii

1

l

I

t)

1

I

$

I

I

.

T

o

”

“

t

l

l

k

$

I,

,

,

r

1

,

t

r

C

1

C

)

1

@

{

:

1

$

BI

b

I1

1118$1f

~

1

I

1

I

1

1

I

1

1

[

1

11

-

.

-

1

I

$

I

f

I

s

$

I

I

1

1

.

.

L

J

J

.

L

J

l

L

1

$

1

$

11

t

1

I

18

-

-

.

.

+

1I

M



40.00

35.00

30.00

20.00

15.00

—dai1507

— Linear [dai1507)

I

10.00

~

1

4

I

I

1

17-Nov

22-Nov

27-Nov

02-Dec

Figure 42. Dial reading and linear

07-Dec

12-Dee 17-De

regression.





Dial

reading

I * I

2 HtmrS / 2 Ifays

i

/’%”i

%’

, ,

-- . --- - - - ---

‘---: --- :

5 Htmrs ./ 5 Hays

. 3.3DIICI. ----: ----

----;----

1+--- ----: ----

----; ----

.___; ----

1 I I

I I

I I >

—-- . - - - - -- - - --

,

I

6 HmrS / & Ilays

. 0.3000-. ----:---___~-_---_.

.__-; --- ---------

__-. -;--.” ---; .-_.

t

,

,

n

,

1

11--- ---:----

----? ---

-;

. =-- -- - . --

i

7

Hcmrs }

7 Ilays

0.3000- ____;---- -__-;----- .__-;---- __-j--_- -_-_;___ .---;----

-----

i?

0.- ---:”-

1

--’- .

--------

-- - .- .. - - ----

L.-__,

---

,

:

i ,

Ilimsm

•1

32.357

MIGRIIN

llli11506 R2

-2. E-03

MM/YmR

l17i11506 .R5

0.07225

MM/YElm

DA11506 R6

0,0’7244

MMmm

llFJ1506 R7

0.06903

MM?fwm

17-Nov-94 00:00:00

12hridiv

24-Ncw- 14 00:00:00

Figure 14. Corrctsinn rate vs TimE base

97512 Corrosion Monitoring and Control in Refinery Process Units (51300-97512-Sg)

Documents