Appendix 8 Case study Assessing Operating limits …efcweb.org/efcweb_media/Downloads/EFC+WP15/Minutes4a_App8_11_WP15.pdfCase study Assessing Operating limits for C-0.5Mo steel in

Minutes of EFC WP15 Corrosion in the Refinery Industry 9 March 2004

Appendix 8

Case study

Assessing Operating limits for C-0.5Mo steel in

high temperature H2 service

Hennie de Bruyn Statoil

Assessing operating limits forC-0,5Mo steel in high temperature

H2 service (case study)

Hennie de BruynEFC WP15 – 9 March 2004

Case study:C-0,5Mo in high temperature H2

Coker distillate hydrotreating unit– Commissioned in 1975

– C-0,5Mo steel pipingReactor feed lineReactor outlet lineDesulph. line (by-pass line)

– Satisfactory performanceNo history of cracking (HTHA)

– Unit optimisation/debottleneckingIncreased operating temperatureIncreased H2 partial pressure


35(20)

300(35)

51(20)

29(20)

265(35)

46(20)

By-pass(presulph)

20350511631544Reactor outlet

35300512926546Reactor feed

pH2(bar)

Temp.(ºC)

Pressure(bar)

pH2(bar)

Temp.(ºC)

Pressure(bar)

ProposedCurrentLine

Is this acceptable?

What about potential damage by high temperature hydrogen?


High temperature hydrogen attack (HTHA)– Dissociation of hydrogen

H2 = 2HThermally driven

– Carbide reaction4H + MC = CH4 + M

– Results in:Decarburization (high temperatures; lower pH2)

– Cr-Mo steelsInternal fissuring/cracking (higher pH2)

– CS; C-0,5Mo; Cr-Mo steels


Nelson curves (G.A Nelson - Shell Development Co)– Collection of experience with high temperature hydrogen

– Temperature-hydrogen partial pressure curves showing experience with different steels

– Paper presented to API in 1949

– Various updates 1950’s & early 1960’s.

– Update in 1965More curvesC-0,5Mo raised by 50ºF


API 941 (Steels for Hydrogen Services at Elevated Temperatures and Pressures in Petroleum Refineries and Petrochemical Plants)

– 1st Edition: July 1970Based on Nelson’s 1965 curves + 2,25Cr-1Mo

– 2nd Edition: June 1977Lowered curve for C-0,5Mo

– Failures in conditions around the curve

– 3rd Edition: May 1983More C-0,5Mo failures (catalytic reforming units)

– 4th Edition: April 1990More industry failures of C-0,5Mo1977 curve removed & presented separatelyCaution on use of C-0,5Mo



2035016315Reactor outlet

3530029265Reactor feed

pH2(bar)

Temp.(ºC)

pH2(bar)

Temp.(ºC)Line


Disadvantage of curves in API 941– Does not indicate ageing effects– HTHA is also time dependent

Single parameter PV was developed relate time, pH2, temperature– API 581 Appendix I: HTHA Technical Module

– PV = log(pH2) + 3,09 x 10-4 (T) (log(t) +14)

pH2 – hydrogen partial pressure in kgf/cm2

(1 kgf/cm2 = 14,2 psia)T – temperature in K (K = C + 273)t – time (age) in hours


Carbon & low alloy steel susceptibility to HTHA (API 581)

PV [ 6,366,36 < PV [ 6,536,45 < PV [ 6,53PV m 6,532,25Cr-1Mo

PV [ 5,835,83 < PV [ 5,925,92 < PV [ 6,00PV m 6,001,25Cr–0,5Mo

PV [ 5,435,43 < PV [ 5,515,51 < PV [ 5,60PV m 5,60C-0,5Mo(normalised)

PV [ 4,784,78 < PV [ 4,874,87 PV [ 4,95PV m 4,95C-0,5Mo(annealed)

PV [ 4,534,53 < PV [ 4,614,61 < PV [ 4,70PV m 4,70Carbon steel

No susceptibility

Low susceptibility

Medium susceptibility

High susceptibility

Critical PV factorsMaterial


Susceptibility to HTHA: C - 0,5Mo steel (annealed)Reactor inlet (Hydrogen partial pressure = 35 bar)

260.0

270.0

280.0

290.0

300.0

310.0

320.0

330.0

340.0

350.0

360.0

0 5 10 15 20 25 30 35 40

Plant Age (years)

Tem

pera

ture

(°C)

High Suceptibility Medium Susceptibility Low Susceptibility Current plant age


Susceptibility to HTHA: C - 0,5Mo steel (normalised)Reactor Inlet (Hydrogen partial pressure = 35 bar)

360.0

370.0

380.0

390.0

400.0

410.0

420.0

430.0

440.0

450.0

460.0

0 5 10 15 20 25 30 35 40

Plant Age (years)

Tem

pera

ture

(°C

)

High Susceptibility Medium Susceptibility Low Susceptibility Current plant age


Susceptibility to HTHA: C - 0,5Mo steel (annealed)Reactor Outlet (Hydrogen partial pressure = 20 bar)

280.0

290.0

300.0

310.0

320.0

330.0

340.0

350.0

360.0

370.0

380.0

0 5 10 15 20 25 30 35 40

Plant Age (years)

Tem

pera

ture

(°C

)



Susceptibility to HTHA: C - 0,5Mo steel (normalised)Reactor Outlet (Hydrogen partial pressure = 20 bar)

400.0

410.0

420.0

430.0

440.0

450.0

460.0

470.0

480.0

490.0

500.0

0 5 10 15 20 25 30 35 40

Plant Age (years)

Tem

pera

ture

(°C

)



Conclusions– Uncertaincies about metallurgical condition

Annealed vs normalisedPiping systems: unlikely to be “normalised”

– Reactor inletRestrict temperature to 275ºC @ pH2 = 35 barLow susceptibility

– Reactor outletRestrict temperature to 315ºC @ pH2 = 20 barLow susceptibility

– Change of materials (1,25Cr-0,5Mo) if higher temperatures are required


Appendix 9

Naphtenic acid corrosion

Studies at Eni Technologie

Roberto Riva Eni

1

San Donato Milanese, 9 March 2004

Naphthenic acid corrosion of 9%Cr 1%Mo steel

Roberto Riva, Giovanna Gabetta, Marino Tolomio, Pietro Bruni

• Opportunity crude oils are cheap, but potentially corrosive becausethey may contain naphthenic acids and a high concentration ofsulphur.

• Chemical analysis alone cannot predict their corrosivity; still, the totalacid number (TAN) and the content of sulphur are key parameters.

• There is a subtle interaction between naphthenic acids and sulphur,which can be described by a simple model.

At a fixed concentration of naphthenic acids the model predicts 1) naphthenic acid corrosion at low S content 2) inhibition of naphthenic acid corrosion at intermediate S content 3) sulphidic corrosion at high S content.

Background

2

• Testing the model that describes the interaction betweennaphthenic acids and sulphur.

• Developing a laboratory test capable of predicting thecorrosivity of opportunity crudes, which could become aplanning tool for refineries. We focus on furnace tubes.

Aims of the experimental work ofEnitecnologie

• Rotating specimen (9% Cr, 1% Mo) in a sealed autoclave at 340°C.

• Two kinds of oil: 1) Ural crude oil (1.5% wt sulphur, TAN=0.2) 2) lubricant base oil (3.8ppm wt sulphur, TAN=0.16).

• Acidity is increased through addition of a mixture of syntheticnaphthenic acids (Fluka 70340, average molecular weight 230).

• Evaluation of the rate of corrosion through weight loss after cleaning.

• Evaluation of the weight of corrosion product deposited on the surfaceof the metal: difference between weight after test and after cleaning.

Experimental procedure

3

Equal concentration of naphthenic acids(TAN=4.9), but different sulphur content

High content of sulphur Low content of sulphur

Reportedly, a different appearance corresponds to the attack of eithercorrosive agent:- Sulphidic corrosion leads to the formation of sulphidic corrosion scales.- No surface corrosion products after naphthenic acid corrosion.

Weight of corrosion productsHigh sulphur content

Average weight of corrosionproducts (9 tests withvarious TAN values)

7.8 mg/cm2

Low sulphur contentAverage weight of corrosion

products (5 tests withvarious TAN values)

1.1 mg/cm2

Corrosion rate

0,000,200,400,600,801,001,20

0 5 10 15 20

total acid number (mg KOH/g oil)

corr

osio

n ra

te (m

m/y

ear)

High sulphur(Ural)

low sulphur(DW05)

Poli. (lowsulphur(DW05))

4

Naphthenic Acid Corrosion Index (NACI)Craig H.L., 1995

NACI = [corrosion rate (mpy)]/[weight of corrosion product (mg/cm2)]NACI TYPE OF CORROSION< 10 sulphidation or, perhaps, oxidation10 to 100 moderate naphthenic attack, perhaps inhibited by sulphidation> 100 severe naphthenic attack

NACI index

0,00

10,00

20,00

30,00

40,00

50,00

0 5 10 15 20

total acid number (mgKOH/g oil)

NACI

inde

xUral (high sulphurcontent)DW05 (low sulphurcontent)Poli. (DW05 (lowsulphur content)) Threshold of

naphthenicacid corrosion

Discussion of the experimental results

The experimental results are roughly in agreement with the model givenin the literature for the interaction between sulphur and naphthenic

acids. For example:- High sulphur content and TAN=10 lead to sulphidic corrosion- Low sulphur content and TAN=10 lead to naphthenic acid corrosion.

BUT

A very high concentration of naphthenic acids (TAN 10) is necessary forthe naphthenic acid corrosion to start. In contrast, far lower

concentrations can give rise to naphthenic acid corrosion in refineryplants.

5

Pitfalls of sealed autoclave tests

The pressure reached in sealed autoclave tests is probably too high(20 to 30 bar). The compounds containing sulphur decomposegradually during the test, the concentration of dissolved H2S risesconsequently and appears to favour sulphidic corrosion.

• The experiments have confirmed the interaction betweensulphur and naphthenic acids. The results are roughly inagreement with the model given in the literature.

• Very high concentrations of naphthenic acids (TAN 10) wererequired for the onset naphthenic acid corrosion. Suchbehaviour is surprising even for 9% Cr-1% Mo steel, which isquite resistant to naphthenic acids . Far lower concentrations(TAN 0.5) are potentially corrosive in refinery plants. Theexperimental set-up probably needs modifying, if plantconditions are to be reproduced.

Conclusions

6

Alternative autoclave tests

According to the literature, a closer agreement between laboratorytests and plant experience could be obtained with either of thefollowing pieces of equipment:

• Pressure relief valve and condenser: at 3 to 7 bar and threedays’ exposure corrosion rates are much closer to those experiencedon refinery plants. (Kane & Cayard, 1998)

• Refreshed rotating cylinder apparatus: low fluid consumption( 10 litres per 24 hours), high pressure (69 bar) (Pritchard A.M. etal., 2001).


Appendix 10


A neural network approach

S. Trasatti University of Milan

1

Naphtenic acid corrosion: a neuralnetwork approach

Stefano P. Trasatti, Francesco MazzaDip.to di Chimica Fisica ed Elettrochimica, Università di Milano

Via C.Golgi 19 - 20133 Milano

EFC meeting 2004 - Enitecnologie 8/9 March 2004

Neural NetworkMathematical model simulating neuronal activity of the human brain.It’s an Artificial Intelligence System enables us to embed structuredhuman knowledge into workable algorithms. NN learns the “weights”

of the correlation between input and output data apparently notconnected by any model (training step)

Target: to predict output from input not used in the training step and toobtain information on the involved mechanism

2

Neural network structure:Multilayer Perceptron

Transfer functions typically used by NNnodes and neurons

3

Neural Network features• Advantages

– Adaptability to phenomena which are changing– High level og robustness (tolerance limit)– Non-linear calculation tool

• Limitations

– Close box tool (no mechanistic understanding of the processbeing modelled)

– High number of data for training step– Data filtering and codification

Neural Network applications• Recognition of characters and images• Analysis of spectra (Raman, NMR, IR, etc.)• Petrochemical industry (rock properties, well productivity,

seismic images)• Noise filtering• Chemical process optimization (recognition offailure

cases)• Loan calculation (banks)• Medical diagnosis (electroencephalogram analysis)

4

Applications of NNs to corrosion (1)

• H.Smets, W.F.L.Bogaerts – NNs to predict the SCC of304ss in near-neutral solutions as a function of Cl-, O2and T

• Urquidi-Macdonald et al. – NN for predicting the numberand depth of pits in heat exchangers

• D.C.Silverman, E.M.Rosen – NN combined with anexpert system to predict the type of corrosion frompolarization curves

• S.P.Trasatti and F.Mazza – NN to study the initiation andpropagation step in crevice corrosion of SS and relatedalloys in near neutral chloride containing media

Applications of NNs to corrosion (2)

• Ramamurthy et al. – NN to analyse impedancedata for automobile paint finishes subjected tostone impacts

• S.Nesic, M.Vrhovac – NN to predict corrosion ofsteel in CO2-containing solutions

• A.Turnbull et al. – NNs to predict SSCC ofDuplex ss

• T.F.Barton et al. – NN to identify pitting andrevice spectra in electrochemical noise

5


Fe + 2RCOOH = Fe(RCOO)2 + H2Fe + H2S = FeS + H2

Fe(RCOO)2 + H2S = FeS + 2 RCOOH

TAN

Flow Rate

Temperature

%S

Crude oil type

Two phase flow

Metallurgy

23,1318,71,12,357231698

270,451,51,12,357231698

2620,010,030,47,22871550

4250,010,030,47,224,571550

13,40,20,250,42,924669662

1,82,0617,40204801500

320,60,010,030204801560

460,010,0304,560,11500

3,30,010,030,10,744801500

1,20,1818,590,129,2740,91450

0,32,0217,460,18,2930,85,5560

1591,089,030,123,1650,820560

351,089,030,113,2820,86,9560

19,41,089,030,18,8512001400

9,21,089,030,146,116401375

42,20,587,090,117,78201455

24,40,587,090,198201455

15,40,587,090,118,112001400

9,40,010,030,11660,940600

0,60,010,030,312100,9170

2,6012,430,311520,940330

0,7012,430,312100,9170

Corr.Rate (mpy)%Mo%Cr%STANExposure Time (h)Flow rate

(m/s)P (atm)T (°F)

6

Variables

OutputProcess4250Corrosion Rate (mpy)

InputPlant4.640Mo_%

InputPlant200.01Cr_%

InputCrude oil4.170S_%

InputCrude oil46.10TAN

InputProcess7686Exposure Time (hours)

InputProcess70Flow Rate (m/s)

InputProcess691Pressure (atm)

InputProcess72570Temperature (°F)

useVariable typeMax.Min.

Parameter

Distribuzione delle frequenze della variabile "Exposure Tim e"

76.7

21.3

1.0 0.5 0.0 0.0 0.0 0.50

10

20

30

40

50

60

70

80

90

100 200 300 400 500 600 700 800

Classi Exposure Time [h]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze della variabile "Temperatura"

0.5 0.0 1.0 0.0

16.8

27.0

38.1

4.7

11.4

0.50

5

10

15

20

25

30

35

40

45

80 160 240 320 400 480 560 640 720 800Classi di temperatura [°F]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze della variabile "Pressione"

89.4

7.40.0 3.2

0

10

20

30

40

50

60

70

80

90

100

20 40 60 80Classi Pressione [Atm]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze della variabile "Flow Rate"

88.6

0.26.4

0.2 3.2 1.20

102030405060708090

100

1.2 2.4 3.6 4.8 6 7.2

Classi di Flow Rate [m /sec]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze della variabile "Corrosion Rate"

78.5

9.96.2 2.7 1.2 0.5 0.2 0.5 0.0 0.2

0

10

20

30

40

50

60

70

80

90

45 90 135 180 225 270 315 360 405 450

Classi di Corrosion Rate [mpy]

Freq

uenz

a de

i cam

pion

i [%

]

7

Data distribution (%)

Distribuzione delle frequenze della variabile "TAN"

43.1

13.617.6

11.18.4

6.2

0

5

10

15

20

2530

35

40

45

50

8 16 24 32 40 48Classi di TAN

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze del "Zolfo in percentuale"

90.6

1.2 4.2 0.5 1.2 0.5 0.2 0.0 0.2 1.20

10

20

30

40

50

60

70

80

90

100

0.45 0.9 1.35 1.8 2.25 2.7 3.15 3.6 4.05 4.5Classi di Zolfo [%]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze del "Cromo in percentuale"

44.8

1.25.0

7.212.1

0.0

12.617.1

05

1015

2025

3035

4045

50

2.5 5 7.5 10 12.5 15 17.5 20Classi di Cromo [%]

Freq

uenz

a de

i cam

pion

i [%

]

Distribuzione delle frequenze del "Molibdeno in percentuale"

68.6

6.2 7.20.0

11.9

0.0 2.7 2.0 0.0 0.70

10

20

30

40

50

60

70

80

0.45 0.9 1.35 1.8 2.25 2.7 3.15 3.6 4.05 4.5

Classi di Molidbeno [%]

Freq

uenz

a de

i cam

pion

i [%

]

Normalizing range

0.90.14250Corr.Rate (mpy)

0.90.14.40.1%Mo

0.90.1180.1%Cr

0.90.14.10.1%S

0.90.1470.5TAN

0.90.11506Exposure Time (hours)

0.90.170.09FlowRate (m/s)

0.90.1691Pressure (atm)

0.90.1750390Temperature (°F)

Output max.Output min.Input max.Input min.Parameter

8

Main component analysis

1005.5481.8206S_Perc

94.458.5132.7927Cr_Perc

85.9410.143.3278Mo_Perc

75.7910.263.3643FloRate

65.5413.484.4214ExpTime

52.0613.634.4711Temp

38.4215.765.1682Press

22.6622.667.4335TAN

Sum% % Value Ranking in input recordVariable

NN for all materialsCorrosion Rate [mpy]

R2 = 0.63

0

100

200

300

400

500

0 100 200 300 400 500

Experimental

Cal

cula

ted

Corrosion Rate [mpy]

0

50

100

150

200

250

300

350

400

450

1 10 19 28 37 46 55 64 73 82 91 100

109

118

127

136

145

154

163

172

181

190

199

Sample number

Cor

rosi

on

Rate

[m

py] t_CorrSpeed

CorrSpeed

Input layer 8

Hidden layer 1 9

Hidden layer 2 3

Output layer 1

9

NN for Carbon SteelCorrosion Rate [mpy]

R2 = 0.76

0

100

200

300

400

0 100 200 300 400

Experimental

Cal

cula

ted


0

50

100

150

200

250

300

350

400

450

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

106

111

116

Sample nuimber

Cor

rosi

on

Rate

[mp

y] t_CorrSpeed

CorrSpeed

Input layer 8

Hidden layer 1 9

Hidden layer 2 5

Output layer 1

NN for Stainless SteelCorrosion Rate [mpy]

R2 = 0.58

0

100

200

300

0 100 200 300

Experimental

Cal

cula

ted


0

50

100

150

200

250

300

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73 76 79 82

Sample nuimber

Cor

ros

ion

Rat

e [m

py] t_CorrSpeed

CorrSpeed

Input layer 8

Hidden layer 1 7

Hidden layer 2 3

Output layer 1

10

Validation stepCorrosion Rate [mpy]

R2= 0.68

0

100

200

300

400

0 100 200 300 400

Experimental

Cal

cula

ted


0

50

100

150

200

250

300

350

400

450

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101 106

Sample number

Cor

rosi

on R

ate

[mpy

]

t_CorrSpeed

CorrSpeed

Input layer 8

Hidden layer 1 9

Hidden layer 2 5

Output layer 1

Validation step

0,010,030,10da 2,00 a 10,0048,000,001,00400-650

%Mo%Cr%STANExpTime(hours)

FlowRate(m/s)

Press.(atm)

Temp.(°F )

010203040

50607080

1 2 3 4 5 6 7 8 9 10 11

TAN

Cor

rosi

on R

ate

(mpy

)

11

Validation Step

57.435.4180.114.8168.01.920.0644.0

61.5540.5180.114.8168.01.925.0644.0

46.1313180.19.9168.01.919.0644.0

29.732.4180.15.4168.01.916.0644.0

34.053.9180.15.4108.01.918.0644.0

26.4526.4180.19.8163.01.937.0644.0

50.3315.3180.110180.01.916.0644.0

25.0837.4180.14.9104.01.925.0644.0

31.8619.7180.14.9108.01.918.0644.0

33.9321.6180.14.996.01.918.0644.0

21.7336.6180.14.9108.01.928.0644.0

45.8429.5180.14.984.03.220.0644.0

26.0724.4180.11.272.03.222.0644.0

20.137.5180.11.2228.03.219.0644.0

Corr.Ratepredetta

Corr.Rate(mpy)

%Mo%Cr%STANExp.Time(ore)

FlowRate(m/s)

Press.(atm)

Temp.(°F

)


Appendix 11

Nace Italia 2004 Conference

sion in the Refinery Industry 9 March 2004

ItaliaSection

The NACE Italy , section of NACE INTERNATIONAL,organizes on the occasion of the recognition in Genoa whatcapital European of the culture 2004, his 2 ^ annualconference.How custom, such conference is established to introduce, todiscuss, to divulge, the connected problems to thetechnology of the corrosion.The involvement of peoples that work with prevention andcontrol of the corrosion is combined to the exhibition ofproducts and services from the working Societies in thesector.The papers interest all the sectors of the technology of thecorrosion involving her thematic to it correlated (Research,Corrosion, Corrosion Control, Coating, CathodicProtection, NDE, Metallurgy, Materials and Services).Continuous evolution of technology calls us to give anexhaustive contribution : scientific, university, productionand commercial, without forgetting the normative andinstitutional aspects.The partecipation to the Conference becomes essential factorknowledge for deepened of this technology and remains theprincipal way to update the technicians that operated, tovaried levels, in the different areas of the industry.The Firms that present products and/or services have thepossibility to intend and to introduce theirs last novelties intechnological terms, doing grow the market and contributingso to an effectiveness cultural popularization.

CORROGenoa

Jolly HotNov

Final papers ca

Abstracts d Papers dead

The languagecan be in engThe “abstractspapers must befollowing addrpresented them

Istituto ItaliLungobisag Alberto

Tel. 010Fax 010e-mail :

Franco Tel. 010Fax 010e-mail :

“ISTITUTO I

presents

SION 2004 - ITALIA

el Marina - Porto Anticoember 25 – 26, 2004

n be even as “extended abstract”

eadline : May 31, 2004line : September 30, 2004 is italian but the oral presentationlish.”, “extended abstracts” and complete in .doc format and sent it at underesses so like the Firms that want products or services.

ano della Saldaturano Istria 15 – 16141 Genova Lauro / 8341354 / 8341353 [email protected] / 8341372 / 8367780 [email protected]

with sponsorship ofTALIANO della SALDATURA”

Minutes of EFC WP15 Corro

Organizing CommitteeChairman Nace ItaliaCarmelo AielloEni Group – Raffineria di GelaGela (CL)Tel: 0933 842485Email: [email protected]

TreasurerMario CelantPipeteamVia Resistenza 220070 - Vizzolo Predabissi (MI)Tel: +39 (0)2 98232820Email: [email protected]

Vice-ChairmanAlberto LauroIstituto Italiano della SaldaturaLungobisagno Istria, 1516141 GenovaTel: 010 8341354E-mail: [email protected]

SecretaryMassimo CornagoENI E & PVia Emilia, 1 20097 S. Donato Milanese (MI)Tel: 02 52063405E-mail: [email protected]

CALL FOR PAPER

AND

mailto:[email protected]

mailto:[email protected]

Documents

Appendix 8 Case study Assessing Operating limits …efcweb.org/efcweb_media/Downloads/EFC+WP15/Minutes4a_App8_11_WP15.pdfCase study Assessing Operating limits for C-0.5Mo steel in