Energy Procedia 51 ( 2014 ) 135 – 147
ScienceDirect
1876-6102 © 2013 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer-review under responsibility of SINTEF Energi AS doi: 10.1016/j.egypro.2014.07.015
Fi
Abst This CO2-boile based study perfo follo corro © 20 Selec Keyw 1. In Im focu and corro * E-m7th Tr
ireside c
G.
Stein-tract work encompa -free, oxyfuel o er materials. Six d superalloy, w y: S-lean bitum ormed by two l wed by laborat osion resistance 013 The Author ction and peer-r words: oxyfuel co ntroduction mportant rese using on proce optimization osion at varyi Corresponding amail address: stein
rondheim C
corrosion
Brzozowsk
aInstitute ofasses the evalu operated power x alloys, varyin were selected fo
minous coal fro aboratories usin tory exposures e, but also revea rs. Published by review under re ombustion, depos earch activitie ess feasibility of the combu ing temperatur author. Tel.: +49 7 n-brzozowska@i
CCS Confer
n during
ka
a*, R. N
f Combustion and bS uation of depos r plants. The aim ng from state-o or the corrosion om Indonesia ng varying met and isothermal aled some differ y Elsevier Ltd. esponsibility of sit, corrosion es on oxyfue y, its potential ustion process res during oxy711 685 67762; f ifk.uni-stuttgart.d
ence, TCCS
oxyfuel
SO
2
co
orling
b, P.
d Power Plant Te Swerea KIMAB, Ssit and material m of the study f-the-art ferritic n tests. The imp
and S-rich bitu thodologies: ex l laboratory exp rences in the co f SINTEF Energ l coal combu l contribution s. There have yfuel combust fax: +49 711 685 de OR maier@ifk
S-7, June
5-l combu
ontents
. Viklund
b echnology (IFK), Stockholm, Swede l related limita was to determ c-martensitic an pact on corrosio tuminous coal xposures in a co posures only. T orrosion produc gi AS. ustion are in n to CO2 redubeen few det tion, and pape
63491 k.uni-stuttgart.de
-6 2013, Tro
stion con
, J. Maier
a Univ. of Stuttgar en ations on boiler mine the reliabil nd austenitic Fe on of oxy-firing from Venezue ombustion test The tests resultcts formed. progress thr uction through tailed investig ers reviewing
ondheim, N
nsidering
a, G. Scheff
t, Germany r performance a ity of advanced e-based steels t g of two fuels i la. Fireside co rig using coole in a similar ran oughout the h separation an gations regard the oxyfuel prNorway
ng variou
ffknecht
aand the lifetim d and conventi to an advanced is considered in orrosion tests w ed corrosion pro nking of the all
world, prima and sequestrat ding deposits process frequen
us
me of onal d Ni-n the were obes oys’ arily tion, and ntly© 2013 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
indicate high corrosion risk due to a high concentration of corrosive gases such as SO2/SO3, (HCl, H2O and CO2)
[1]-[3].
Fireside corrosion observed in a boiler is a net result of the synergy effect between the alloy, the flue gas and the deposit. Most of the studies on corrosion are performed in laboratory set-ups due to cost limitations; however, it is very difficult to include all the conditions observed in a real boiler during laboratory studies. A majority of the studies run worldwide concentrate either only on the corrosive influence of the gas atmosphere, thus neglecting the interactions with the fly ash deposit, or consider only the artificial fly ashes in the performed tests. Applied deposits often represent a very corrosive environment offering a quick alloy screening but no information on a possible tube life expectancy after being installed in a boiler operating at specific conditions. Many studies deal with flat metal coupons, thus neglecting the influence of tube geometry.
In the presented study, real fly ash from oxy-combustion tests of two hard-coal qualities were used. The gas blends used in the tests depict the real combustion conditions of both coals. Two different test procedures were used to study the usability of advanced and conventional boiler materials and to see to what extent differences in test procedures might affect the resulting corrosion behavior of selected alloys. Both studies were focused on the estimation of corrosion performance of selected materials and the evaluation of deposit and material related limitations on boiler performance and lifetime of CO2-lean, oxyfuel operated power plants.
Nomenclature
abs. absolute
aus. austenitic steel(s) Bal. balanced
BSE back-scattered electron image [Cr] chromium content in wt% EDS energy-dispersive spectrometry fer.-mar. ferritic-martensitic alloy GS ground surface
ICP-OES inductively coupled plasma optical emission spectroscopy k constant referring to oxide character
lee leeward side, side hidden from direct flow of combustion and/or flue gas
luv windward side, side exposed directly towards the flow of combustion and/or flue gas PR S-rich Venezuelian Palma Rejo coal
RFCS Research Fund for Coal and Steel
S sulfur (used as S-lean and S-rich while describing coals and test conditions) SEB sulfur-lean Indonesian Sebuku Coal
SEM scanning electron microscopy SiC-paper abrasive grinding paper TIG-welding tungsten inert gas welding vol% volume percentage w with
WDS wavelength-dispersive spectrometry wt% weight percentage
w/o without
2. Experimental
2.1. Tested materials
The main focus of the corrosion tests was on austenitic steels with varying chromium content. These are: 304L, 153MA, 253MA and 310S (see Table 1). All the mentioned austenitic stainless steels were TIG-welded tubes. Additionally, a martensitic alloy T92 and a nickel-based grade 617 were selected for fireside corrosion tests [4]. Table 1. Composition of the tested alloys (wt%).
Alloy Cr Ni Fe Other
617 22.0 Bal. (54.7) 1.1 C (0.006), Si (0.09), Mn, P, S, Mo, Co, Al, Pb
310S 25.3 19.1 Bal. C (0.046), Si (0.059), Mn, P, S, Co, Cu, N, Nb
253MA 21.0 11.1 Bal. C (0.086), Si (1,6), Mn, P, S, Co, Cu, N, Nb, Ce
153MA 18.5 9.2 Bal. C (0.053), Si (1,16), Mn, P, S, Co, Cu, N, Nb, Ce
304L 18.2 8.2 Bal. C (0.022), Si (0.4), Mn, P, S, N
T92 8.8 0.2 Bal. C (0.1), Si (0.24), Mn, P, S, Mo, V, B, N, W, Cb
2.2. Corrosion tests
The oxyfuel combustion tests were performed using two bituminous coals: a sulfur-lean Indonesian Sebuku and a sulfur-rich Venezuelian Palma Rejo coal. The firing tests were conducted in FoSper, 3 MWth combustion test rig of Enel as described earlier by Cumbo [5] and Stein-Brzozowska [4]. The gas atmosphere composition and fly ash collected during the firing tests were used subsequently as an input for the long term-corrosion tests performed at the laboratories.
At one laboratory (IFK), the samples were pre-exposed on cooled corrosion probes positioned in the combustion chamber of the technical scale test rig to form the first oxide scale and original initial deposit layers (see Table 2 for deposit composition). The short-term pre-exposure (see Figure 2) was followed by a long-term iso-thermal laboratory exposure (see Figure 3). At the other laboratory (Swerea KIMAB), the samples were covered with a deposit slurry followed by a conventional long-term iso-thermal laboratory exposure without any pre-exposure (Figure 1).
The first methodology allows for cost-effective studies of long-term corrosion behavior of samples, which have formed their initial protective oxide scale during real combustion exposure conditions including a temperature gradient across the tube wall. Moreover, the influences of the tube geometry and deposit formation in real combustion conditions are also considered at relatively low cost. Both laboratory test set-ups consisted of horizontal tube furnaces supplied continuously with a dedicated gas blend. The two test temperatures 650°C and 580°C were defined.
For tests without pre-exposure, the samples consisted of circle segments cut from tubes (Figure 1). The TIG-welded alloys 304L, 310S and 253MA were prepared and exposed in two different ways: in the ground surface condition where the weld was not part of the specimen and in the as-received condition including the weld. Additionally, duplicate specimens of T92 and Alloy 617 were exposed with a surface ground to 600 mesh SiC paper. For tests with pre-exposure, whole tube rings with as-delivered surface condition were used allowing observation of leeward- and windward sides of the tube (see Figure 3). Before being exposed, all the specimens were thoroughly cleaned with alcohol and acetone.
During the laboratory exposures, with samples pre-exposed in the test rig, the deposits (Table 2) were renewed every 200 h by spreading new layers of fresh fly ash originating from oxyfuel combustion of S-lean and S-rich coals respectively over the existing initial deposit layer (see Figure 3). For the other test series, after removal of “old” deposit layer, the samples were coated weekly with a fresh fly ash slurry layer generated using the same fly ash as in the case of the pre-exposed tests (Figure 1). For this purpose, during the cooling sequence for tests
with to a pure Tabl Al2 CaO Fe2 K2O Mg Na2 P2O SiO TiO SO3 Figu Figu the c I so t atm inco a) a) hout pre-expo avoid condens e nitrogen in o le 2. Composition O3, wt% O, wt% O3, wt% O, wt% gO, wt% 2O, wt% O5, wt% O2, wt% O2, wt% 3, wt% ure 1. a) Specimen ure 2. a) Rings on combustion cham In total, eight that they mim mosphere was
omplete comb
osure, the gas sation on the order to elimin n of initial deposi S-lean depos 27.5 4.15 3.53 1.37 1.64 0.864 0.721 56.9 2.51 3.77
n cut to the right
corrosion probe mber.
exposure seri micked measu
chosen inclu bustion and air
b)
composition w samples. Dur nate any possi it layer. sit S-rich 14.06 6.43 5.59 1.74 5.24 0.64 0.15 58.92 0.74 6.49 dimensions, b) Sp
before the exposu
ies were run; s ured combust uding CO and r ingress, resp
b)
was deliberate ring the tests w
ible reactions
h deposit 6
2
pecimen covered
ure, b) and after t
see Table 3 fo tion condition d N2 to simu
pectively. c)
ely not altered with pre-expo below the tem
d with a deposit, c
the exposure in
or a test summ ns. For the te ulate the effec
d, except that osed specimen mperature rang c) Specimen place Figure 3. Pre-subsequent 95 exposure. mary. The SO2 ests without p ct of non-equ
the water con ns, the sample ge of interest. ed in the crucible -exposed ring 617 50 h of laboratory 2 and H2O con pre-exposure, uilibrium cond
ntent was low es were coole e 7 after y ntents were va a more com ditions caused wered ed in aried mplex d by
Table 3. Corrosion test conditions during laboratory exposures Test Temp. [°C] Deposit coal type O2 [vol%] CO2 [vol%] H2O [vol%] SO2 [vol%] CO [vol%] N2 [vol%] Exposure time [h] w/o pre-exposure PR 580°C 580°C S-rich 3 66.7 15 0.3 0.01 15 960 w/o pre-exposure PR 650°C 650°C S-rich 3 66.7 15 0.3 0.01 15 960 w/o pre-exposure SEB 580°C 580°C S-lean 3 66.9 15 0.1 0.01 15 960 w/o pre-exposure SEB 650°C 650°C S-lean 3 66.9 15 0.1 0.01 15 960 w pre-exposure PR 580°C 580°C S-rich 3 86 11 0.3 - - 950* w pre-exposure PR 650°C 650°C S-rich 3 86 11 0.3 - - 950* w pre-exposure SEB 580°C 580°C S-lean 3 82 15 0.1 - - 950* w pre-exposure SEB 650°C 650°C S-lean 3 82 15 0.1 - - 950*
* only laboratory tests considered not including the 50-hour pre-exposure time
2.3. Analytics
After completed exposures, the metal specimens were investigated by light optical microscopy and scanning electron microscopy (SEM) with energy- and wavelength-dispersive X-ray spectroscopy (EDS, WDS) to analyze the morphology and composition of the corrosion products. The corrosion behavior of the materials was compared by measuring the total metal loss after completed exposures and by determining the corrosion rate with respect to 1000 h estimated as indicated in formula (1) and measured with the help of SEM-images of the corrosion products. The total metal loss corresponds to the reduction in the load-bearing tube section and the corrosion rate presented and discussed in section 3.1 always relates to 1000 h.
Total metal loss = max observed depth of internal corrosion + k* (max observed external oxide scale) (1) Precise analysis requires small differences in the k-values used, because the oxide types vary significantly (iron oxide, chromia, spinel, mixed spinels, spinels mixed with single oxides), resulting in varying oxide-to-metal density ratios and cation-to-anion ratios. For the tests without pre-oxidation, the k-values were determined to be in the range 0.43-0.47 depending on the alloy. For simplicity, an average constant k = 0.5 was used for the tests with pre-exposure, since this is common industrial practice and the error is almost negligible.
Some of the original deposits on the corrosion probe rings from the pre-exposures were collected and analyzed by ICP-OES and XRD directly after removal from the combustion chamber and before the laboratory tests. The procedure was repeated after the long-term corrosion tests were completed. The change in composition of the deposits is discussed in section 3.2.
3. R 3.1. 3.1. A by mar exp test Figu for T with temp ferrit L exp Wit look deb atta orig Nev exte prec [6]. wer the Results and d Corrosion 1. Corrosion As expected, w arrows in Fi rtensitic (indi posure conditio
rig. The main
ure 4. Corrosion ra T92 and 617 from h ground surface. perature. (see arro tic-martensitic m Looking at Fi posed in the te th increasing king at the co batable since t ack along grain
ginates partly vertheless, the ent during th cipitation of a Important to re exposed, m observed inte discussion rate with increased igure 4 and icated by gre ons, higher co n differences b
ates during tests a m the tests w/o pre Higher corrosion ows). Austenitic a material (fer.-mar.)
gure 6 and Fi est rig, a clea chromium co orrosion rates the alloys 310
n boundaries from the tube e irregular pit he test series alumina along note is that th meaning that ap ergranular atta d temperature Figure 5). M ey frames in orrosion rates between the st at S-lean conditio e-exposure relate n rate with increas alloys (aus.) perfo ).
igure 7 where ar dependence
ntent, the cor determined at 0S and 617 do observed at b e manufacturin s present alon without pre g the grains w he non-pre-exp pprox. 30 µm ack reaches. e, higher corro Moreover, aus Figure 4 and at 1000 h we tudies were no
ons. The results to samples sing
orm better than
e the corrosion e between chr rrosion rate de t the pre-expo o not follow t oth grade 310 ng process can ng the metal s e-exposure. In was reported e posed specime of the origina osion rates at stenitic alloys d Figure 5). ere not always oticed in the b Figure 5. Corro for T92 and 61 with ground su increasing tem perform better
n rates are dis romium conte ecreases as th osed species (F
the trend. Hig 0S and 617 (se annot be exclu surface of 310 n the case o earlier in fire ens of grade 6 al tube surfac 1000 h were s performed However, rel s evident for behavior of gr
osion rate during 17 from the tests w urface. As during mperature (see arro than ferritic-mar
splayed for th ent and alloy he rule of the Figure 8 and F gher corrosion
ee Figure 10 a uded after anal 0S (see Figur of Alloy 617, side corrosion 617 underwen e material wa noticed in bo generally bet lating alloy p the samples p ades 310S and tests at S-rich co w/o pre-exposure S-lean tests: high ows); and austeni rtensitic material ( e specimens t performance thumb claims Figure 9), this n rates at both and Figure 11 lyzing the pre e 14) were no , intergranula n studies unde t machining b as removed, w oth studies (sh tter than ferr performance pre-exposed in
d 617.
onditions. The res e relate to sample her corrosion rate itic alloys (aus.)
(fer.-mar.).
that were not can be obser s. However, w s rule seems t h tubes are du 1). That this at e-exposed spe ot detected to ar oxidation der air-combus
before the sam which is as dee hown ritic-with n the sults s e with pre-rved. when to be ue to ttack cies. o this with stion mples ep as
Figur expos to sam with i Figur expos depen better Figur bound re 6. Corrosion ra sure. The results f mples with groun increasing chrom
re 8. Corrosion ra sure. Original tub ndence between in r corrosion resista
re 10. Grade 310S daries (courtesy o
ates during tests a for T92 from the nd surface. Improv mium content.
ates during tests a be surfaces consid ncreasing chromi ance. S, pre-exposed PR of Outokupmu). at S-lean condition tests w/o pre-exp ved corrosion per
at S-lean condition dered. Debatable
ium content in all
R 650°C. Attack a ns w/o pre-posure relate rformance ns with pre-direct loy matrix and
along grain Figure 7. Corro exposure. The r relate to sample performance wi Figure 9. Corro exposure. Origi dependence bet alloy matrix and
Figure 11. Grad boundaries (cou
osion rate during t results for T92 an es with ground su ith increasing chr
osion rate during t inal tube surfaces tween increasing d better corrosion
de 617, pre-expos urtesy of Outokup
tests at S-rich con nd 617 from the te urface. Improved c romium (plus nick
tests at S-rich con considered. Deb chromium (plus n n resistance.
sed PR 650°C. At pmu).
nditions w/o pre-ests w/o pre-expo
corrosion kel) content.
nditions with pre-batable direct
nickel) content in
ttack along grain osure
3.1. F an i lean sulf duri Figu Exte gas p O wer at 5 in S on t of f disp An Ast Nev low D of s 14) corr cap 2. Sulphur ind
For the test se increased corr n specimens ( fides were det ing S-rich test
ure 12. Sulphides ent of sulfides inc phase and deposit
On the other h re varied, surp 580°C, 310S, 6 S-lean vs. 11 %
the overall cor further resear played higher increased ten eman [9]. Mo vertheless, oxi w and thus hav During the S-r sulfur in the o . On the wind rosion produc acity of the de duced corrosi eries, without rosion rate fo (compare Figu tected at the ts and with inc
on T92, exposed creases with temp
t. hand, tests wi prisingly show 617 at 580°C) % in S-rich). W rrosion rate th rch. In the stu corrosion rate ndency in al ore authors ag idation rates o ve not been def rich tests with oxide scale wa dward side, (l cts during S-r eposits (see se
ion and oxidat
pre-exposure or some alloys ure 6 vs. Figu oxide-metal i creasing temp d without pre-expo perature and S-con
ith pre-exposu wed slightly hi ) during S-lean Whether or no han a threefold udies of Hol es at higher w loy oxidation gree that wate observed on a fined as probl h pre-exposur as observed m luv) where th rich tests (see ection 3.2). tion in a combust s displaying e ure 7: at 650 nterface of se perature (see F osure. ntent in the ure (Figure 8 a igher corrosio n tests, which ot a small incr d difference in comb et al. [ water contents n with elevat er plays a sign austenitic alloy lematic for a f re, no sulfides mostly on the he deposit was Figure 14). T tion environm even double a °C: 310S, 25 everal alloys Figure 12 and Figure 13. Sulph Extent of sulfide the gas phase and
and Figure 9) on rates for mo h were charact rease in H2O c n SO2 concen [7], T91 (sim in conditions ted water con
nificant role i ys during the full scale boile s were noticed leeward side, s regularly re This is possib
ment, increased and higher cor 3MA and T9 (T92, 304, 25 Figure 13). hides on 304L, exp s increases with t d deposit. ), where not o ost specimens terized by a 4% content (15 % ntration (0.1 % milar to T92) s simulating ox
ntent was als in oxide form studies descri er operation un d in the corros where the de efreshed, barel bly an effect d SO2 concent rrosion rates t 92). Chromium 53MA), and a
posed without pre temperature and S only SO2but a (T92 see Figu %-abs.higher H % vs. 11 %) ha % vs. 0.3 %) sh and 347TP ( xyfuel combu so observed b mation and its ibed in this pa nder oxyfuel c sion product. eposit was not ly any sulfur of the observ ntration resulte than respectiv m and manga at higher amo e-exposure. S-content in also H2O cont ure 15, 304L H2O content (1 as a greater im hould be a sub (similar to 30 ustion atmosph by Piron [8] stability [8]-[ aper are relati conditions. A slight pres t renewed (Fi was found in ved sulfur-bin ed in ve S-anese ounts tents only 15 % mpact bject 04L) here. and [13]. ively ence igure n the nding
Figur Barel refres T cont oxid Figur oxide furthe re 14. Alloy 310S ly any sulfur foun shment).
T92 is characte taining less iro dation selectiv re 15. Typical oxi e scale formed by er down an intern S: sulphur in corro nd on the windsid erized by form on but more c ve to Cr, reach
ide structure obse y Fe2O3, below lyi
nal oxidation sele
osion products: m de (left column), w mation of dou chromium oxid hing almost 20 erved at T92. Upp ing a mixed oxide ctive to Cr.
more S found on th where deposit wa
uble-layered th des and includ 0 µm, was dete
per row: S-rich w e layer containing
the leeward side ( as regularly refres
hick oxide-sca ding Cr-rich i ected.
w pre-exposure 58 g less iron, more
(right column) wh shed. Some S fou
ale: outer Fe2O slands (Figure 80°C, bottom row chromium oxides here no deposit re nd at leeward sid O3-dominated e 15). Additio w: S-lean w pre-ex s and including C efreshment took p de (no deposit d and inner mi onally, an inte xposure 580°C. O Cr-rich islands, place. ixed rnal Outer
I oxid oxid Som in [4 F oxid spec mea men Figu T form app As exp N 253 3.1. S as i to a mac thei tube A mat in o 3.2. D com ash sulp Irregular oxid des characteri des prevail on me buckling e 4]. For grade 617 de scale is stil cimens. The m asurement, is ntioned in sect ure 16. S-rich, 617 The ground su m, whereas a proximately 30 ground 617 sa plained by the No signs of a 3MA specimen 3. Outlook
Since the mas ntergranular o apply visual a chining. Altho ir impact on t e, which is als Although the terials need to oxyfuel fired b Deposit During both fi mpleted expos as reported i phate is detect
des are observ ized by an ou n 304L. Moreo ffect was noti 7, both studies ll very thin, ex main differenc the irregularl tion 3.1.1 and 7 pre-exposed at urface of 617 all the four pr
0 µm and conf amples do not original tube accelerated co ns, showing th ss change mea oxidation, thu analytical tech ough the use o the corrosion so relevant to present study o be confirmed boilers.
firing tests, the sures in comb
in [4] and [5 ted on the coo
ved along the uter hematite over, 304L is iced occasiona show negligi xcepting some ce between bo ly occurring in d depicted in F 580°C: intergranu 7 samples test re-exposed 6 firmed by oth t reveal any fa surface being orrosion attac hat the corrosi
asurements of us potentially d hniques while of ground surf of the observ consideration y contained rea d in long-term e primary dep bustion test rig ]. Although t oled surfaces [ surfaces of 15 and inner ch characterized ally on 304L a ible oxide scal
e external oxi oth laboratory ntergranular o Figure 16). ular oxidation wi ted without pr 17 specimens er studies on o ailures, the int
removed duri ck were notic ion resistance f the exposed delivering a w determining c faces allows fo ed alloys, it d n during corros asonably long m field tests to
posit layers for g revealed mu the earth alka
4]. 53MA, 253M hromia layer, d by internal at at 580°C but le at lower tem ide nodules ob y measurement oxidation acco
ith alumina precip
re-exposure d s indicated pr original 617 tu ternal oxidatio ring the grindi
ed at the pos of the welds w
specimens do wrong impress corrosion rate for a better com
does not repre sion studies. g exposures, th o allow for a s rmed collecte uch higher su ali content is MA and 304L. although at ttack along gr not detected a mperatures. W bserved occas ts, which is no ompanied by pitation.
did not show resence of in ube material i on observed at ng process. sition of the was comparab o not depict si sion of the all es of original mparison of d esent the actu he trends in c safe predictio
ed from corros lfur content c low in the co
All three allo lower temper rain boundarie at 650°C as re With increasing ionally on the ot depicted by the aluminum any presence tergranular ox n the as-receiv t pre-exposed welds on the ble with the ba
ignificant corr loy’s performa surfaces not i ifferent alloy al surface con corrosion beha n of the tube
sion probe rin compared to o oals used, cal
oys tend to cr ratures the m es, similarly 3 eported previo g temperature e pre-exposed y the mass cha m precipitation e of this corro oxidation reac ived condition d 617 rings ma e 304L, 310S ase metal. rosion forms ance, it is adv influenced by compositions ndition of the avior of the te materials life ngs (Table 2) other collected lcium magnes reate mixed 10S. ously e, the d 617 ange n (as osion ching n [6]. ay be and such vised y any s and real ested time after d fly sium
In again clear sulfu depo Figur the te T of a toge Alth corro spec of S to ac T that no e obse had How n order to sup n after the lab rly shown usi ur from the s osits after test
x by a f x by a f
re 17. S-enrichme est rig, S-content i
The XRD analy anhydrite and ther with K a hough a simult osive role in h cimens. Moreo -rich tests, con ccelerate the c The sulfur enri this feature ha eutectic melts ervation that,s
not been ren wever, further
pport the evalu boratory tests ing a standard sulfurous spec termination, s factor 3 in dep factor of 1.2 in
ent in deposit ove increases in depo yses confirme calcium-mag and O in the f taneous prese hot corrosion over, an incre nfirms that th corrosion proc ichment of de as a partly pro of fly ash an sulfur among ewed during observations uation of the of pre-expose d ICP-OES m cies present in
sulfur was fou posits of S-rich n deposits of S
er exposure time a osits during expos
ed the sulfur e gnesium sulph form of an S-ence of K, S a type II [14], ease of anhydr e fly ash captu cess since they eposits over tim
otective role a nd alloying el the corrosion exposure, and are necessary influence of d ed samples had method and is n the gas pha und to be high h coal ash S-lean coal ash
at 580°C and 650 sures in the labora
enrichment of hate. Addition -rich rim surr and O points t still no signs rite and calciu ures sulfurous y remain stabl me indicates t against sulfur-lements were n products, sul d only during . deposit on fir d been comple shown in Fig ase and in co her: h. 0°C. Compared to atory. the deposits. nally, looking rounding parti to the presenc s of this type um magnesium s species from le in the obser the sulfur-bind -induced firesi found on the ulfur was obse g the S-rich t eside corrosio eted. An enric gure 17. In su omparison wi o initial S-content The element w g at SEM-W icles of a Ca-ce of potassium of corrosion w m sulphate ov m the gas phas
rved condition ding capacity ide corrosion, e samples. Th erved mostly ests in the ca
on, the deposi chment of sulf ummary, the d th the initial in initial deposit was mostly pr DS element m Mg-rich core m sulphate, kn were observed ver time, espe
e. The two ph ns. of fly ash. It rather than th his is further s at positions w ase of pre-exp
its were analy fur over time w deposits captu S-content in t layer as collecte resent in the fo maps, S appe e (see Figure
nown for its v d in the analy ecially in the c hases do not se can be presum he opposite, si supported by where the dep
posed specime yzed was ured the ed in form ears 18). very yzed case eem med ince the posit ens.
Figu (K) a 4. C T allo chro test coo A the loss perf not Cr surf prob resi A enri sulf indu foun obs test F an i duri allo ure 18. Deposit on and oxygen (O) a
Conclusions The influence oys: a ferritic– omium conten s were perfor oled corrosion Although som laboratories, s. The highest formed substa pre-exposed s content and t faces. The ve bably a cons istance [16]. H Analysis of th ichment of sul fur-binding ca uced fireside nd on the sam erved mostly s. Further obs For the test se increased corr
ing S-rich test oys exposed at
n pre-exposed in are found in the ri
of combustio –martensitic g nt 304L (18% rmed by two l probes follow me discrepanci the two type t corrosion ra antially better species includ the corrosion ery good perfo
equence of th However more
he deposits p lfur over time apacity of fly corrosion, rat mples. This is at positions w servations are eries without p rosion rate. Su ts. Surprisingl t S-lean tests S-reach condition im as well. on of two hard grade T92, a %Cr), 153MA laboratories u wed by laborat ies are observ s of tests resu ate was obser r. Improved pe ding some spe
rate could no formance of tw
heir high Si e tests are sugg performed afte e, reaching a fa ash. It can be ther than the o further suppo where the dep suggested wit pre-exposure ulphides were ly, during the characterized ns 310S. Typical d coals on fire nickel-based A (18.5%Cr), using varying tory exposure ved in the mor
ult in very sim ved on the fe erformance w ecimens charac ot be establish wo of the aus content, whic gested to conf er completed factor of three. e speculated th opposite, sinc orted by the ob posit had not b th varying min in a combust found in the c
tests with pre by a slightly
S-rim around cal
eside corrosio d superalloy 6 253MA (21% methodologie s and isotherm rphology and milar alloy ra erritic-martens was noticed w cterized by a hed in case o stenitic steels ch is added t firm this. tests with sa . The sulfur en hat this featur e no eutectic bservation tha been renewed neral systems tion environm corrosion prod e-exposed spe higher H2O c
lcium and magne
on performanc 617 and four %Cr) and 310 es: exposures mal laboratory composition anking and co sitic steel, wh ith increasing ground surfac of pre-exposed s (253MA 21 to improve th amples pre-ex nrichment of d re has a partly melts of fly a at, among the d during expos . ment, increased ducts of sever ecies, slightly content than r
sium-rich fly ash
ce was studied austenitic all S (25%Cr). F in a combust y exposures on of corrosion omparable dim
hile the tested g Cr content in
ce; a direct rel d specimens w %Cr and 153 he high temp xposed in the deposits over y protective ro ash and alloyi
corrosion pro sure and only d SO2 concent
ral alloys and a thicker oxide espective S-ri
h particle. Potassiu
d on the follow loys with var Fireside corro tion test rig u nly. products betw mensions of m d austenitic al n the cases of lation between with original 3MA 18.5%C perature oxida test rig reve time indicate ole against su ing elements w oducts, sulfur y during the S-ntration resulte at higher amo es were notice ich tests. Whe
um wing rying osion using ween metal lloys f the n the tube Cr) is ation ealed s the ulfur-were was -rich ed in ounts d on ether
or not a small increase in H2O content (15 % vs. 11 %) in the oxyfuel combustion atmosphere characterized by
high partial pressure of CO2 has a greater impact on the overall corrosion rate than a threefold difference in SO2
concentration (0.1 % vs. 0.3 %) should be a subject of further research.
It should be taken into account that mass change measurements of the exposed specimens do not show significant corrosion forms such as intergranular oxidation, thus occasionally delivering wrong impressions of an alloy’s performance. Therefore, the determination of corrosion rates is suggested to be either verified or performed mainly using the visual observations with known micro-analytical techniques. Moreover, although the use of ground surfaces allows for a better comparison of different alloy compositions and their impact on corrosion of the observed alloys, it does not reveal the actual condition of the applied tube, which is also relevant to corrosion studies.
Acknowledgements
The above described work was performed within the frame of the OxyCorr Project which was partly funded by the RFCS Research Program of the European Commission (RFCR-CT-2009-00005). The authors are very grateful to the European Commission for the support.
The authors would like to express their gratitude to Dr. Rachel Pettersson (Outokumpu Stainless) for her valuable contribution to the work and to Outokumpu Stainless for providing the TIG-welded austenitic stainless steel tubes. All the unnamed employees of IFK-KWF, IFRF and Enel are acknowledged for supporting the combustion tests. Besides, the authors owe special thanks to Dipl.-Ing. Maike Johnson for her very helpful remarks.
References
[1] Scheffknecht G, Al-Makhadmeh L, Schnell U, Maier J. Oxy-fuel coal combustion – A review of the current state-of-the-art. Int J
Greenhouse Gas Control 2011, 5: S16-S35.
[2] Toftegaard M, Brix J, Jensen P. Oxy-fuel combustion of solid fuels. Prog Energy Combust 2010; 36: 581 - 625.
[3] Stanger R, Wall T. Sulphur impacts during pulverised coal combustion in oxyfuel technology for carbon capture and storage. Prog Energy
Combust 2011; 37: 69 – 88.
[4] Stein-Brzozowska G, Maier J, Scheffknecht G, Cumbo D, Masci S, Tosi E, Corraggio G, Faleni M, Biasci L. Fireside corrosion of applied and modern superheater-alloys under oxyfuel conditions. GHGT-11. Tokyo, Japan 18-22.11.2012.
[5] Cumbo D, Rossi N, Tosi E, Stein-Brzozowska G, Miller E, Maier J, Scheffknecht G, Viklund P, Coraggio G, Tognotti L. Corrosion of candidate superheater materials during oxyfuel conditions - Pilot plant and laboratory investigations. OCC2. Yeppon, Australia. 12-16.09.2011.
[6] Stein-Brzozowska G, Flórez DM, Maier J, Scheffknecht G. Nickel-base superalloys fpr ultra-supercritical coal-fired power plants: Fireside corrosion. Laboratory studies and power plant exposures. Fuel 2013; 108: 521-533.
[7] Gordon R. Holcomb, Tylczak J, Meier GH, Lutz BS, Yanar NM, Pettit FS, Zhu J, Wise A, Laughlin DE, and Sridhar S, Oxy-Combustion Environment Characterization: Fire-and Steam-Side Corrosion in Advanced Combustion, National Energy Technology Laboratory-In-house Research, 2012.
[8] Piron-Abellan J, Olszewski T, Meier GH, Singheiser L, Quadakkers WJ. The oxidation behavior of the 9% Cr-steel P92 in CO2- and H2
O-rich gases relevant to oxyfuel environments. Int J Mater Res 2010: 287-299.
[9] Asteman. H, Svensson J, Johansson LG. Effect of Water-Vapor-Induced Cr Vaporization on the Oxidation of Austenitic Stainless Steels at 700 and 900°C. J Electrochem Soc 2004, 151:141-150.
[10] Huenert D, Schulz W, Kranzmann A. Corrosion of steels in H2O-CO2 atmospheres at temperatures between 500°C and 700°C Federal Institute of Materials Research and Testing, Berlin; September 8–11, 2008. P R E P R I N T – ICPWS XV.
[11] Montgomery M, Hjörnhede A, Gerhardt A. Short-term corrosion testing in a burner rig with oxyfuel and conventional firing. Nice, France; September 6 – 10, 2009; 5-20.
[12] Natesan K, Rink DL, Zeng Z. Materials performance of structural alloys in simulated oxyfuel environments. 24th Annual Conference on Fossil Energy Materials. Pittsburgh, USA; May 25-27, 2010.
[13] Stein-Brzozowska G, Maier J, Scheffknecht G. Deposition Behavior and Superheater Corrosion under Coal Fired Oxyfuel Conditions, Special Workshop on Oxyfuel Combustion Addressing SO2/SO3/Hg and Corrosion Issues. London, Jan. 25-26, 2011.
[14] Khanna A.S. Introduction to high temperature oxidation and corrosion. 1st edition, ASM International; 2002, ISBN: 0-87170-762-4. [15] Surman PL, Castle JE. Gas phase transport in the oxidation of Fe and steel. Corros Sci 1969; 9: 771-772.
[16] High Temperature Stainless Steels. Outokumpu MA grades. http://www.outokumpu.com/SiteCollectionDocuments/Austenitic-High-Temperature-153MA-253MA-stainless-brochure.pdf last visited 31.05.2013.
