VOLUME 43 ISSUE 1

of Achievements in Materials and Manufacturing Engineering

November 2010

Evaluation of reformer tubes degradation after long term operation

J. abanowski*
Gdask University of Technology, Faculty of Mechanical Engineering, 11/12 Narutowicza st., 80-233 Gdask, Poland * Corresponding author: E-mail address: jlabanow@pg.gda.pl
Received 10.09.2010; published in revised form 01.11.2010

Properties
ABSTRaCT
Purpose of this paper is to show the effect of long-term service at elevated temperatures on microstructural changes of cast steel reformer tubes made of the alloy IN-519 (24%Cr, 24%Ni, Nb). The relationship between mechanical properties and microstructure degradation is discussed. Design/methodology/approach: Investigations were performed on five tubes taken from ammonia reformer furnace. Tubes worked at temperature of 880C and 3.2 MPa pressure. The operation time varied from 24000 to 95000 hours. Microstructure of the material has been examined with the use of light optical microscopy and scanning electron microscopy (SEM). The energy dispersive X-ray analysis (EDX) was used for phase identification. The mechanical properties have been evaluated in static tensile tests at room temperature. Findings: Metallographic investigations of IN-519 cast steel catalytic tubes show that structural changes occur in non-monotonic way with operating time at elevated temperatures. For this reason the evaluation of tube degradation can not be based only on the microstructure examinations. The presence of a good correlation between mechanical properties of the catalytic tubes and service conditions, i.e. temperature, time, and internal stress was confirmed. Research limitations/implications: The surface condition of the tubes walls like carburization and oxidation is not considered in this study. Practical implications: The method for assessing the current degradation level and for predicting residual lifetime of creep-resistant tubes has been established. The method uses elongation values as the main factor characterizing structural degradation of catalytic tubes. It respects all degradation mechanisms involved in tubes degradation, and can give reliable information on reformer tube condition through almost whole period of operation. Originality/value: Microstructure transformations occur in the IN-519 cast steel tubes during long-term operation in the reformer furnace were revealed and described Keywords: Creep-resistance; Structure degradation; Reformer tubes; IN-519 cast steel Reference to this paper should be given in the following way: J. abanowski, Evaluation of reformer tubes degradation after long term operation, Journal of Achievements in Materials and Manufacturing Engineering 43/1 (2010) 244-251.

1.  Introduction 1. Introduction
Catalytic tubes are important parts of reformer units at ammonia, methanol, hydrogen and gas process plants. They are the most expensive parts of reformer equipment. A steam

reforming process converts hydrocarbons into mixture of hydrogen, carbon monoxide and dioxide. Chemical reactions proceed at a temperature range of 800-900C and under pressure of 3-4 MPa. These severe working conditions cause a structural damage of tubes. The most widely used austenitic cast steel tubes are susceptible to creep at working temperature, and it is essential

244

Research paper

Copyright by International OCSCO World Press. All rights reserved. 2010

Properties

to be able to identify and locate such damage at its early stages. The ammonia plant users postulate strong needs to develop reliable methods for inspection of tubes degradation and for realistic prediction of theirs residual life. Uncertainty about residual life frequently leads to conservative operation, beneath of nominal productivity of reformer units or premature replacement of the tubes [1-3]. Prediction of residual lifetime of creep-resistant tubes and pipelines has been based on investigations performed during operation, including creep tests, tube deformation or deformation velocity measurements, metallographic examinations, mechanical properties tests (static, impact and fatigue) and physical properties examinations (radiological, ultrasonic, magnetic and electric properties measurements). The basic criterion for the evaluation of the suitability of creep-resistant material for future operation is creep strength, which depends on the material structure, and always decreases during the tubes lifetime. Creep tests are expensive and long term, so there are undertaken attempts to use more cost-effective methods instead of creep tests that also give useful and significant information [4-11]. Reliable criterion describing the amount of degradation in creep resistant reformer tubes has not been established so far. Conventional non-destructive testing (NDT) techniques such as eddy current and ultrasonic, currently applied to reformer tubes, are geared to detect creep damage in the form of internal cracking and are useful in the last stage of operation in reformer furnaces. Carburisation and oxidation of austenitic tubes surfaces create ferromagnetic layers. This phenomenon is frequently used as an indicator of tubes degradation. Measurements of magnetic permeability and low magnetic fields can give satisfactory information on effective tube wall thickness but cannot indicate creep failures unless micro cracks appear [2,3]. The most recent techniques employ internal or external tube diameter growth as a material degradation indicator. Whenever catalytic tubes are operating under pressure in their creep temperature range, their diameters will increase in time. The laser mapping method referred as laser profilometry allows to measure and quantify the tubes internal diameter [12]. The lasermapping probe, inserted in a tube, can transmit several thousand diameter readings down its length. Recorded diameter increases in the range of 1-6% are potential indicators of certain stages of creep. This simple in principle and very convenient method does not assume other degradation mechanisms such as carburisation or oxidation of internal or external tube surfaces. The method described in [13] assumes detecting of creep damage in catalytic tubes by two complementary examinations: ultrasounds testing and diameter measurements. The material structure, and the presence of voids can be assessed by means of ultrasounds testing. External diameter measurements of tubes deformation describe quantitatively creep damage. Information is added to database and converted in the life-assessment program, which calculates the probability of crack initiation and tube failure. Methods mentioned above are still in progress but due to the complexity of factors influencing the life of the catalytic tubes, current life prediction methods are limited and are not able to detect and quantify reliably of the damage existing in the tubes. The structural criterion characterizing tube degradation is very convenient but it can be accepted only when the structure changes

continuously with the working time. This method, for example, is widely used for assessing quantitatively degradation of steel operated at elevated temperatures in power plants. Similar correlation can be observed for other metallic alloys working under creep conditions. Many investigations on heat resistant austenitic alloys reveal microstructure changes occurred during their operation in reformer units [1,13-15] but there are not any reliable structural factor proposed as degradation indicator so far. In this study, the centrifugal cast tubes made of alloy IN-519 after various working times are considered. The effects of longterm service at elevated temperatures on microstructural changes have been studied. The relationship between operation time and mechanical properties is discussed. The method for assessing the degradation level of tubes based on mechanical properties examinations at room temperature was proposed.

2.  Experimental 2. Experimental
Investigations were performed on five tubes made of IN 519 cast steel taken from ammonia reformer furnace. The tubes worked at temperature of 880C and under 3.2 MPa pressure. Samples were taken below 4 m from the inlet end of the tubes where temperature during operation is stable. Tested tubes worked for varying lengths of time from 24000 to 95000 hours. The tubes of a total length 12 meters (Fig. 1) were assembled together by welding from the 3-meter segments, so the chemical composition of the alloy could slightly differ along the tube.

Fig. 1. Catalytic tube used in reformer unit

Table 1 shows the chemical composition of the samples. Since the new material was not available, the reference structure was detected on the sample PM taken from inlet end of the tube, where operational temperature did not exceed 540 C. Microstructure of the alloy has been examined with the use of light optical microscopy and by scanning electron microscopy (SEM) Chemical compositions of various phases have been examined by energy dispersive X-ray analysis (EDX). Samples were mechanically polished and etched in Murakami reagent (30 g K3Fe(CN)6, 30 g KOH, 60 ml water). The mechanical properties at room temperature have been determined in static tensile tests. Metallographic examinations were performed on the cross sections at the area near 1/3 tube wall thickness from the inner surface. The surface condition of the tubes walls (carburisation, oxidation) is not discussed in this paper.

READING DIRECT: www.journalamme.org

245

Journal of Achievements in Materials and Manufacturing Engineering

Volume 43 Issue 1 November 2010

Table 1. Chemical composition of tested IN-519 cast steel tubes Sample designation PM 24-1 24-2 30-1 30-2 44 72-1 72-2 95 Distance from the tube inlet m 0.2 6.2 6.5 6.2 6.5 4.0 4.0 8.0 5.0 Working time h 24000 24000 30000 30000 44000 72000 72000 95000 Chemical composition wt.% C 0.297 0.351 0.352 0.309 0.435 0.286 0.354 0.318 0.331 Si 0.598 0.62 0.72 0.79 0.71 0.731 0.833 0.86 0.659 Mn 0.557 0.53 0.52 0.68 0.61 0.518 0.467 0.365 0.464 Cr 23.3 23.42 23.65 23.79 23.40 23.4 24.7 23.9 23.87 Ni 26.0 25.59 25.36 25.17 24.66 25.7 24.9 24.9 2 6.17 Nb 1.38 1.46 1.57 1.47 1.8 1.47 1.50 1.46 1.63 S 0.005 0.003 0.002 0.002 0.001 0.003 0.009 0.012 0.006 P 0.029 0.021 0.020 0.024 0.008 0.023 0.025 0.023 0.013

3. Results and discussion 2. Results and discussion

3.1.  Metallographic examinations 2.1. Metallographic examinations
The non-degraded (PM) specimen showed the dendritic columnar austenite grains located perpendicularly to tube walls. The structure consists austenitic matrix with a proportion of primary inter-dendritic eutectic niobium carbide and chromium carbide. These massive lamellar carbides closely resemble pearlite. The thin semicontinous network of eutectic carbides can also be seen in the iterdendric areas, Fig.2.

The original morphology of eutectic carbides has been modified during the long time of operation at elevated temperatures during which certain coalescence has taken place.

Fig. 2. Non-degradated structure of IN-519 cast steel, specimen PM An EDX analysis showed that some of eutectic carbides are Nb-carbides and others are Cr-carbides. The two types of carbides cannot be distinguished from each other in the optical microscope The microstructures of IN-519 cast steel specimens after various working times are presented in Figs. 3-10.

Fig. 3. Microstructure of IN-519 cast steel tubes after 24000 h operation in reformer furnace, (a) 24-1 and (b) 24-2 specimen

246

Research paper

J. abanowski

Properties

this area of the furnace were the same, the manufacturing technology was also similar, so the structural difference could only be explained through slightly different chemical compositions of tube segments.

Fig. 5. Microstructure of IN-519 cast steel tubes after 44000 h operation in reformer furnace

Fig. 4. Microstructure of IN-519 cast steel tubes after 30000 h operation in reformer furnace, (a) 30-1 and (b) 30-2 specimen The structure of 24-1 and 24-2 samples contains still lamellar eutectic of primary carbides, but the coalescence process has formed not completely continuous network of primary carbides. Secondary carbides (mainly Nb carbides) are observed within the grains, but to a rather varying degree, both in size and amount. In addition to the small and rounded carbides inside the austenite grains, coarser and plate or needle-like precipitates are observed. An EDX analysis revealed that these precipitates have a chemical composition close to the sigma phase. Presence of this phase in the alloy structure was unexpected because IN-519 cast steel has low susceptibility to formation of this embrittling phase in the temperature range of 600-950C. To ensure low susceptibility to sigma phase formation, the chemical composition of the alloy was modified to conform to the upper limit of 1.0% silicon and to maintain the nickel/chromium ratio at unity or above. The amount of sigma phase and degree of carbide coalescence was different in samples 24-1 and 24-2 (Fig. 3). These two samples were taken from different segments of one reformer tube but very close to each other (0.2 m) separated by welding joint. Obviously, the working conditions in

The degree of degradation in the 30-1 and 30-2 specimens (30000 h) was greater than in 24 samples (Fig. 4). The carbide coalescence process is more advanced and quantity of lamellar eutectic decreased, but it still exists in both samples. The grain interior was filled with needle like sigma phase precipitates and dispersed NbC carbides. These two samples were taken from two segments of one reformer tube very close to each other and also in these structures the considerable differences in the amount of secondary phases were observed. The structure of 44 sample (44000 h) consists continuous network of coalesced carbides around austenite grains with only the trace of primary eutectic structure. Very small amounts of sigma phase were detected inside the austenite grains. The similar structure was found in 72-1 sample (72000 h). In this case, the semicontinuous coarse network along austenite grains consists of Cr and Nb carbides. Sigma phase is also present in the structure in the form of blocky precipitates. In 72-2 (Fig. 6b) sample, the corresponding coarse network of coalesced carbides existed but grain interiors were filled with great amounts of needle-like sigma phase precipitates. A small number of creep voids were found in 72 samples structure. The voids were formed at the interface between matrix and primary carbides or between matrix and the blocky sigma. The voids were round and rather small (up to 8 m in diameter) and limited in number. They are scattered with no tendency to lining. The structure of 95-1 was similar to 72-2 sample (Fig. 7), but unetched structure revealed a great number of creep voids. The voids were located mainly on the inside surface of the tube and at the midwall. Voids are becoming aligned along dendrite boundaries. Aligned voids coalesced into fissures (up to 800 m in length) were detected close to the internal surface of the tube (Fig. 8).

Evaluation of reformer tubes degradation after long term operation

247

Journal of Achievements in Materials and Manufacturing Engineering

Volume 43 Issue 1 November 2010

a)

b)

Fig. 6. Microstructure of IN-519 cast steel tubes after 72000 h operation in reformer furnace, (a) 72-1 and (b) 72-2 specimen

c)

Fig. 7. Microstructure of IN-519 cast steel tubes after 95000 h operation in reformer furnace

Fig. 8. Microstructures of IN-519 cast steel tubes. Sample 95, a), b), c) voids aligned along dendrite boundaries

248

Research paper

J. abanowski

Properties

Changes in IN-519 tube microstructure during operation in reformer furnace

matrix

primary and secondary phases

chemical composition modifications and accompanied heterogeneity in solid solution changes in dislocation structure and dislocation density growth changes in austenite grain boundary microstructure due to the carbides coalescence and voids formation

changes in primary inter-dendritic austenite-carbide eutectic secondary carbides precipitation and interactions with primary carbides continuous carbides network formation around austenite grains coalescence of sigma phase inside and around austenite grains

effects of microstructural changes in IN-519 cast steel

deterioration of mechanical properties

decrease in creep resistance microvoids initiation on interphase boundaries voids propagation

voids coalescence into fissures

Fig. 9. Microstructural evolution of IN-519 cast steel tubes during operation time in reformer furnace [2] Metallographic observations revealed that the following transformations took place in the structure during tube exploitation [2,15]: coalescence of primary inter-dendritic austenite-carbide eutectic, formation of continuous network of primary and secondary carbides around austenite grains, sigma phase precipitation in the needle-like form, coalescence of sigma phase to the blocky shape. The amount of primary and secondary phases (NbC, M23C6, ) existing in the structures is presented in Fig. 10.

Evaluation of reformer tubes degradation after long term operation

249

Journal of Achievements in Materials and Manufacturing Engineering

Volume 43 Issue 1 November 2010

A m ount of prim ary and secondary phases, %

30 25 20

a)
600 Y = -113.295 * log(X) + 1653.13 Coef of determination, R = 0.9116

500

10 5 0 0 20000 40000 60000 80000 100000

TS, MPa
400 300

15

10000

Working time, [h]

Working time, [h]

100000

Fig. 10. Total amount of primary and secondary precipitates existed in the IN-519 alloy structure after long term service at elevated temperature This simple quantitative description of the degree of structure degradation shows that structural changes during the operation time are not uniform and change in a non-monotonic manner with operating time. Moreover, the structures of the samples taken from the same reformer tube, exploited in the same conditions also can differ in quantity and form of precipitated phases.

b)
15 13 10

Y = -1.17042 * log(X) + 15.9754 Coef of determination, R= 0.9650

E, [%]

8 5 3

3.2.  Mechanical properties 2.2. Mechanical properties examinations examinations

Excluding application of microstructure factor for degradation level description of reformer tubes, the mechanical properties should be considered. The literature data indicate that mechanical properties obtained in static or dynamic tests at room temperature can not be used as measure of cast steel tube microstructure degradation mainly due to the non-monotonic changes after various operation periods. That theory was confirmed in investigations [4] but only for the first, short period of tubes utilization (up to 20000 h). Further operation can provide linear relationship between specific mechanical property and working conditions. Results of hardness tests and tensile tests performed at room temperature on flat test specimens are shown in Fig. 11. The specimens were taken parallel to longitudinal axis of the tubes and contained the whole tube wall thickness. The time axis at the charts is plotted in logarithmic scale in order to obtain the straight line and estimate correlation between mechanical properties and tube working time. The hardness test results show a large scatter (Fig. 11a) and it seems that this method is not suitable for estimating reformer tube degradation. Results of tensile tests (Fig.11b,c) indicate that the most significant is reduction of plastic properties measured by unit elongation (E) at room temperature. The E values show the best correlation with working time logarithm. Taking this into account, the degree of E values reduction could be useful for determining the tubes degradation level and for estimating its residual lifetime.

0 10000 2 3 4 5 6 7 8

Working time, [h]

9 100000

c)
280 260 Y = -10.8672 * log(X) + 328.185 Coef of determination, R = 0.2118

HB2,5/187,5/15

240 220 200 180 160 10000 2 3 4 5 6 7 8

Working time, [h]

9 100000

Fig. 11. Mechanical properties of IN-519 alloy after long-time service at 880C in reformer furnace, a) tensile strength, b) elongation, c) Brinell hardness

3.3. The method time 2.3. The method for for life life time prediction prediction of reformer tubes

of

reformer tubes

The method employing mechanical property changes can be applied for assessing the current degradation of the tubes material after considering working parameters, e.g. stress and

250

Research paper

J. abanowski

Properties

temperature. Working conditions can be representing in LarsonMiller parameter (PLM). Standard relationship between tubes elongation (E) and PLM was established after analysis of several dozen investigations results taken from available literature. The good correlation of regression equation was obtained. Practical verification of E =F(PLM) relationship was confirmed in examinations performed on IN-519 tubes (Fig. 12). Correlation between mechanical properties and service conditions allows on predicting residual lifetime of creep-resistant tubes.

References References
[1] [2] J. Barcik, Alloys for pyrolitic tubes, University of Silesia, Katowice, 1995. J. abanowski, Assessment of catalytic tubes degradation of reformer units in service, Gdansk University of Technology. Monografy 35, Gdansk, 2003 (in Polish). J. abanowski J, Non-destructive tests for reformer tubes degradation assessment, Chemical Engineering and Equipment 4-5 (2000) 14-18 (in Polish). D. Renowicz, A Hernas, M. Ciela, K Mutwil, Degradation of the cast steel parts working in power plant pipelines, Proceedings of the 14th Scientific International Conference Achievements in Mechanical and Materials Engineering AMME'2006, Gliwice - Wisa, 2006, 112-116. S.B Parks, C.M. Schillmoller, Improve alloy selection for ammonia furnaces, Hydrocarbon Processing 76/10 (1997) 93-98. C.W. Thomas, M. Borshevsky, A.N. Marshall, Assessment of thermal history of niobium modified HP-50 reformer tubes by microstructural methods, Materials Science and Technology 8/10 (1992) 855-861. A. Marek, G. Junak, J. Okrajni, Fatigue life of creep resisting steels under conditions of cyclic mechanical and thermal interactions, Archives of Materials Science and Engineering 40/1 (2009) 37-40. A. Hernas, G. Moskal, K. Rodak, J. Pasternak, Characterisation of properties and microstructural changes of 12% Cr-W steels after long-term service, Journal of Achievements in Materials and Manufacturing Engineering 31/2 (2008) 312-319. J. Okrajni, Thermo-mechanical fatigue conditions of power plant components, Journal of Achievements in Materials and Manufacturing Engineering 33/1 (2009) 53-61. J. Okrajni, K. Mutwil, M. Ciela, Steam pipelines' effort and durability, Journal of Achievements in Materials and Manufacturing Engineering 22/2 (2007) 63-66. R.D. Roberts, P. Tait, Laser profilometry applied to catalyst tubes in steam reformers for ongoing condition monitoring of creep damage, Insight 42/8 (2000) 525-527. M.C. Perez, Integrated approach for life assessment of reformer tubes, Proceedings of the Conference Plant Life Assessment PLAN, Spain, October, 2001. T.L. Silva, I. Le May, Damage accumulation mechanisms in reformer furnace, Proceedings of the Conference Materials Ageing and Component Life Extension CIM, Montreal 1995, 1-10. M.B. Zaghloul, T. Shinoda, R. Tanaka, Relation between structure and creep rupture strength of centrifugally cast HK 40 steel, Transaction of ISIJ 17 (1977) 28-36. J. abanowski, Assessment of structural changes in IN-519 cast steel reformer tubes, Advances in Materials Science 5/2 (2005) 15-22.

[3]

[4]

[5]

[6]

[7]

Fig. 12. Elongation of IN-519 alloy in dependence of working parameters in reformer furnace [4] The method was developed on the base of IN-519 cast steel investigations, and uses elongation values as the main factor characterising structural degradation of catalytic tubes. The proposed method respects all degradation mechanisms involved in tubes degradation, i.e. creep, oxidation, carburisation, thermal fatigue and can give reliable information on reformer tube condition through almost whole period of operation.

[8]

[9]

[10]

[11]

4. Conclusions 3. Conclusions
[12] 1. Investigations of IN-519 cast steel reformer tubes revealed that structural changes occur in non-monotonic way with operating time at elevated temperatures, thus the evaluation of tube degradation cannot be based only on structure appearance. 2. Changes of elongation values for IN-519 cast steel reformer tubes obtained in tensile tests show the best correlation with the working parameters. 3. The method for assessing the current degradation level and for predicting the residual lifetime of creep-resistant tubes based on mechanical properties examinations performed at room temperature was proposed.

[13]

[14]

[15]

Evaluation of reformer tubes degradation after long term operation

251