Document 96BBwmV9ko0dL12zYnLNbGk7

This material has been reproduced with permission of the copyright owner. Further reproduction is prohibited. Perfromonce of Ductile-Iron Pipe in Soils Melvin RomanoH----------------------------------------------------------- ------------------------------------------------------------ A paper presented on /if11. S, 1967, at the Annual Conjeience, Atlantic City, fi'J., by Melvin Romanoff, Research Chemist, Rational Bureau of Standards, Institute for Materials Research, Washington, ll.C. THE field investigations on under ground corrosion, conducted for loss and pitting, as gray cast iron in the same soil environment, and that many years by the National Bureau of the graphite corrosion products of Slandards, have contributed signifi ductile-iron specimens have the same cantly to our present understanding general characteristics and adhesive of the behavior of metals in soils. properties as those of gray cast iron. Daring the 30 years preceding 1952 It was concluded, therefore, that in the data were obtained on 37,000 speci same soil ductile iron and gray iron mens, representing many varieties of not only corrode at about the same metals, alloys, and protective coatings, rate, but that the pattern of corrosion some of which had been exposed up to and the nature of the corrosion prod 17 years at the 12S test locations ucts arc also similar. throughout the United States. There In this art'cle data arc presented, data, reported in part in many differ giving another point on the corrosion ent publications, were all assembled in time curves, after exposure for 8 Xational Bureau of Standards Circu-. years, in soils at five of the six test lar 579.1 sites where specimens of ductile iron In 1958, new specimens were buried were exposed. There are additional at six test sites to obtain additional in specimens, still exposed, awaiting later formation on metals and alloys not com removal. pletely investigated in the earlier tests. Before discussing the results ob Included were specimens of ductile tained on the specimens of ductile iron cast-iron pipe. Carbon steel pipe was in the field tests, it is advi-ablc to re used as a reference material to corre view briefly the data on the perform late the corrosiveness of the soils at ance of gray iron in the earlier tests. the six test sites with that of the soils Present at the 128 XBS test loca at the 128 locations at which corrosion tions are 95 different types of soils. investigations were previously con The chemical and physical projicrtics ducted.1 of these soils cover a wide range. Kx- A progress report was made on the treme acidity fpU 2.C>). as well as performance of ductile iron after ex high alkalinity (pH 10.21. is repre posure of 4 years/ The data pre sented. Electrical resistivity ranges sented in this progress re|>ort indicated from 51 ohm-cm, approximately that that ductile cast iron corrodes at nearly of sea water, to 54,000 ohm-cm, indi the same rate, with respect to weight cating the absence of soluble salts. In CTD000149 646 MELVIN ROMANOFF /"r. .Ill'll one soil the soluble material is almost entirely in the form of sulfates, while in other soils it is in the form of chlo rides or carbonates. Wide differences in aeration are exhibited. Some of the environments arc highly oxidizin';. Others arc definitely reducing. Grey-Iron Data The gray cast-iron materials, buried at 47 test sites in 1922, consisted of several varieties of pipe specimens manufactured by both the ccntrifugat and pit cast processes. Data obtained on these materials after exposure for 12 years in the more corrosive soils and 17 years in the less corrosive soils show that in the same soil environ ment there is no appreciable difference with respect to both weight loss and pitting in thccorrosion of the east-irurmanufactured by the different nieilio;. The data also show that the ci.rn.si,. of gray iron is determined cs-ctnia!: by soil conditions, not bv the conig silinn of the material, and, fortified supplementary data obtained at is .. dilional sites in subsequent vears. i! show that soils differ radically in t!,r. corrosive action on cast iron and m!ncominonly used ferrous metals. Ii one soil the corrosion rate, as indicate.' by weight loss and maximum depth pits, may be constant. In another. may decrease with time. This is illn. trated in Fig. 1, which shows tla changes in weight losses and maxiimi-i pit depths relative to ex]M>surc time for gray iron in four different soils. In general, the rate of weight hor depth of pining in well drained vf. with high electrical resistivities. mi.\. as I lagerslown loam and Snsipichair clay, is high initially, but becomes i' significant after a few years. In - aof this type, an increase in wall (liicl. ness will substantially increase the ser vice life of a pipe. On the other liar! in poorly drained soils with low a sistivitics, such as Acadia clay an I-tkc Charles clay, the rate of rnrr. sion is nearly linear after 1 or 2 >enrIn these soils heavier-wall pipe mu not be justified economically. Ik-chik it is more practical to apply protcctiu coatings, cathodic protection, or Isoth Fig. 1. Weight Loss and Maximum Pit Depth `of Oray Cast Iron Plotted Against Exposure Time Cast-Iron Corrosion Corrosion of cast iron is character ised by conversion of the metal into aadherent layer consisting chiefly of irw oxides and graphite, which incrow in thickness as corrosion progrrw Corrosion by this process is cal!r graphilizatiou, or graphitic cnrr"-u,,r and is caused in part by eleclrnchcmi cal action between the ferritic mx CTD000150 hums TITE rF.HKOItMANCE 647 rig. 2. Corrosion-tl'nc Curves for Gray Cut Iron and Carbon Steel Burled in Hagerstown Loam for 14 Yeara graphitic constituents of cost iron and also, perhaps, by dilTciernes in poten tial produced by contact of the metal with the stiil. An example of this would be differential aeration. The granhitized lover often ran he detected by its dull hl.nl; appearance and by the lack of a metallic ring when the metal is struck bv another metallic object. Nonetheless, because the orig inal sha|)c of the metal is retained, vis ta! observation prior In thorough clean ing and removal of the graphitized corrosion products gives no indication ol the extent of corrosion. For the pur|x>se of corrosion studies, graphilization is considered to lie cor rosion, although it is generally recog nized that the strength of the grapliitized corrosion products is often suffi cient to permit a severely corroded pipe to provide many years of service if low nr moderate pressures arc maintained, and the pipe is not subjected to me chanical disturbances in the surround ing soil. To evaluate the residual strength of iraphitizcd east iron, specimens of 2-in. OD pipe with a wall thickness of J in., that had been removed from five cor rosive soils after exposures of up to 11 years, were subjected to hydraulic pressure tests.1 After the application of hvdraulic pressures, the corrosion products were removed, and the specimens were ex amined for weight loss and pitting. The data show that most ni the speci mens withstood a maximum pressure of 500 psi, although removal of the cor rosion products revealed numerous perforations of various diameters. Gray cast iron is a heterogeneous alloy, chiefly because the carbon that it contains is predominantly present as graphite flake:. The graphite flakes tend to disrupt the continuity of the metal, causing hritilcness and lowsi length. Ductile Iron Ilv contrast ductile iron is manu factured lu- a process that converts the graphite Hake to m-dulcs more or less uniformly distributed throughout the metal matrix. This ductile cast iron, also known as nodular or spheroidal cast iron, is considerably tougher and more ductile than gray iron. In 195R, specimens of 2-in. ductileiron pipe, 12 in. long w ith a w-all thick ness of ] in., were buried in six dif ferent soil environments. Twenty specimens were buried approximately 2-4 ft Mow the ground line at each lest site to allow for removal of one set consisting of four specimens after 1, 2, 4, and fi years, with one set to lie removed still later, depending upon the results of the prior removals. The progress report mentioned above gives data for specimens exposed 1. 2, and 4 years in the six soils r The present article -,'lds R-year data in five of the soils. The sixth site, coastal sand (site E), was washed out hy a CTD000151 648 MELVIN ROMANOFF Jour. AW IVf tidal flood prior to the scheduled re moval of the 4-year specimens. There fore, only 1- and 2-ycar data are presently available for the s]>ccimeirs from this site. Additional specimens were exposed in a nearby site of simi lar soil characteristics from which data will be available in the future. Test Sites The physical and chemical proper ties of the soils arc given in Table 1. Descriptions of the soils at the six test sites follow: 1. Site A. Sagemoor sandy loam is a well-drained alkaline soil with a resistivity of 400 ohtn-cm and a pH of 8.8. It is typical of soils found in vast 0 2 4 6( tiposur. Ttmc-ycjfi Fig:. 4. Hagerstown Loam Weight loss and maximum pit depth c- plotted as a function of time for dial': cast-iron (solid lines) and carbon-nee' (dashed I:::::) pip: b:aicd i:: r, irni'ii loam (Site II) at i.oili Katrn. Me o2* 6 s EipotiKC Tim*-yem Fig. 3. Sagemoor Sandy Loam Weight loss and maximum pit depth arc plotted as a function of lime for ductile cast-iron {solid lines) ami carbon-steel {dashed lines) pipe buried in Satjemoor sandy loam {Site .4) at Toppeuislt, Wash. areas of eastern Washington and Ort gon. Titc site is located on tS Yakima Indian Reservation near To: penish, Wash. The soil is consistr in composition to a depth of at least ft and supports an almndant growth' sage brush. 2. Site B. Hagerstown loam is t well drained soil representative of th well-dcvclo]>c<l soils found in the east tern United States. The site is Icatcd at the Loch Raven Reservoir' the Haltimore water department. T1 soil consists of a brown loam about ' ft deep, underlain liv a mldish-brmv clay extending down 5 ft or inure t a rock base. The soil lias a re-asm r of 5,200 ohm-eni and a pi I of - v Almost all the materials investigate since 1922 in N US soil corrosion tea- CTD000152 ;.wMS PIPE PERFORMANCE 649 have been exposed at this site. There fore, it can serve usefully as a refer<nce site for correlating data obtained Airing the present program with that of earlier tests. 3. Site C. Clay soil is located in a large clay pit on level land at Cape May. N.J. The soil consists of plastic gray clay to a depth of 6 in. This is underlain by gray clay mixed with patches of brown clay to a depth of 12 in. Underneath this is a poorly drained, very heavy plastic clay in which the specimens arc exposed. The soil has a resistivity of 300 olun-cm and a pll of 4.0. 4. Site D. Lakewood sand is a white, loose sand with black streaks in some places. The site is located in a well drained, rolling area at Wildwood, X.J., which is not subject to overflow from the ocean except under unusual Hood conditions. The sand, which sup ports the growth of beach grasses abundantly, lias a pH of 7.3 and a resistivity of 30,000 ohm-cm. 5. Sile F-. Coastal sand is a typical white beach sand with a high content of black sand streaks. It is similar In the I-akcwood sand, except that at this site, on the Two-Mile lleach at Wildwood, N.J., the sand is saturated TAHLE 1 Properties of Soils at Test Sites Site Soil Ixxjtion Internal Drainage of Tint Site A Sagcmoor sandy loam Toppenish, Wash. B IluLcrLloWii !.l111 c Cb> D Lakewood sand :.wj. mu. CapeMav.S. 1. Wildwood, N. j. F. Coastal sand Wildwood, N. J. G Tidal marsh I'atuxcnt, Md. Good Good I'oor Good Poor l'tmr K.-inivitj-* n/iinrw 4(H) 5,200 300 .to.non 55 .ton Pit 8.8 5.8 4.0 7.3 7.1 7.1 TDSt A 7,080 C 14,640 F. 11,020 G II,.180 c* 108 540 302 14(1 Composition of water extract--ppm Mb Na + K (as Na) CO1 23 1,960 0.0 IlCOi 5,002 Set* 216 754 2,242 0.0 tt.o 6,768 329 3,230 0.0 165 2,392 0.0 55 1,133 0.0 1,709 Cl 330 3,529 5.765 3,259 NO, 6 118 31 37 Mg cquivalcnts/HHIji of soil - 0.54 0.19 8.50 0.0 8.20 - 2.7(1 6.18 9.51 0.0 0.0 1.51 2.7(1 13.9 -- 0.7(1 1.35 1(1.2 0.0 0,09 0.0 0.0 * Resistivity determination* marie at the test site with Sln'iiard Cjik*. i Toul dissolved solids, residue dried at 10S*C. 0.-*5 14.0 3.36 3.56 J0.93 0.01 9.Ml j i 10.2 | 9.IK j 0 19 (1.(15 u.oo CTD000153 650 MELVIN KOMAKOFP continuously with salt water. It has a pH of 7.1 and a resistivity of 55 ohm-cin. 6. Site G. Tidal marsh is typical of the poorly drained soils found alum; the Atlantic and Gulf coasts. The site is located along a creek that empties into the Chesapeake Kay at l'anixcnt, Md. The soil is charged with hy drogen sulfide and has a resistivity of 300 ohm-cm and a pi I of 7.1. Gray Iron Nol Exposed Because gray iron was not exposed in these tests, no direct comparison Jour. A WII'J H'cight loss and maximum pit depth co plotted as a function of time for dncti.e cast-iron (solid lines) and carbon-sled (dashed lines) pipe buried in /.abexere sand ot IFitd-onod, A'./. Weigh! loss ami maximum fit depth are platted as a fuuctiou of time for ductile cast-iron (solid lines) and carbon-steel (dashed lines) pipe buried in clay (Site C) at Cape May, NJ. could be made between it and dnetik iron. Comparisons between s]iecinmiof ductile-iron and carbon-steel pipe (1.9 in. OD, 12 in. long, 0.145 in wall thickness), which were exposed t the same sites, were made to evaluVr the behavior of the ductile-iron sped mens. A statistical comparison o: earlier data obtained for specimens e: gray iron and steel, after exposure f 14 years in the fifteen soils, slump1 that no significant differences hciwcer the weight losses of the two materials could be detected.1 It is reasonable tr' assume, therefore, that the same sit" alion exists when gray cast iron and ductile cast iron are compared. Wiy respect to maximum jieiu-lrntioii. ir leu snils. no significant ditfereuerx In'tween the two materials could lie de tected. In five soils pit depth was si;- CTD000154 i.uim PIPE PERFORMANCE 651 nificantly greater for tlio specimens of fray cast-iron. It is also worth noting that data obtained for the many corrosive soils io which both gray cast iron and steel were exposed in the earlier tests 1 show that, despite the greater wall thick ness of cast-iron pi|\ the first perfora tions on both materials, caused by cor rosion, generally were detected alter the same cx)iosorc jtcriod. The curves for weight loss and maximum pit depth plotted against ex posure time for s|<ecimciis of grayiron and steel pii>e exposed in the earlier tests to the moderately corrosive Hagerstown loam fsitc II) are shown in big. 2. The curves show that with respect to weight loss both materials corroded at about the same rate, and that the rale of pitting on the cast iron was greater than that of steel. The weight losses and maximum pit depths plotted against the cx|xisurc time for the specimens of ductile-iron and carbon-steel pipe in five soils are TAItl.l- 2 Average* Loss in Weight and Maximum fit Depth of Curbatt Steel and DuetHe Cast Iran in IHfferettl Sails Urot. Soil TytK- L<*c;<li<>rv A SagCtlKKir saudv loillll TopiKMiUh, Wash. ('iirlxui yrart I | UVrclil i..*- i , ifl U 1.0 2.0 1 4.0 8.0 0.9 2.5 hi' 4.4 B Hagerstown loam l>H-h Haven, Md. 1.0 2.1 4.1 8.1 0.9 1.5 .1.4 c Clay Cape May, -V J. 1.0 1.0 2.0 3.3 5.0 7.6 8.0 42.7 1) takewood tv'imi Wildwood, X. J. t.tt t.4 2.(1 1.8 4.0 2.6 8.0 4.0 E Coastal sand Wildwood, X .1. 1.0 tu 2.0 0.6 G Tidal mar>h i'ntnxcut, Md. 0.9 1.1 Md. 1.K 2.0 3.9 .14..1 7.8 40.0 -- Average ot Omr < * O** 0,1 >|nvirnfti3 [crloful*-d hy corrosion. '*{* null tlilcknej; carbon i*el. I-15 null; ductile (run. 3JO miU. Avttije of three ipeclmcn*, Otic ipectnicn nut recovered. Steel j Dllililr Iron Mux. I'll | WVtjjlit i Mji. I*ii lMOi i Los. 1 mi/> j os ; rih .to i 11.4 Ui 1.7 55 It 71 1.4 45 66 60 niit itt 20 55 60 j 12 30 1211 1.3 2.0 i> 3.1 2.4 4.7 5.`> 14.8} 49 68 81 97 61 108 114t 2311 1>J 0.7 33 1.4 54 2.1 45 2.3 50 44 43 52 15 1.8 35 8 5.8 49 12 14 l.Wit 12Xt 2.7 3.4 1U.4 17.1. 64 ' 56 | 142t I'Mlt CTD000155 6S2 MELVIN ROMANOFF I Jour.AWW/. shown in Pip. 3-7. Each point repre sents an average of the weight losses and maximum pit depths of four sjiecimens. The weight loss and pit depth data from which the curves were plotted arc given in Taldc 2. The condition of the specimens of ductile cast-iron pipe and carbon-steel pipe after exposure for 8 years at the test sites arc shown in Fig. 8 and 9. The curves for both weight loss anti pit depth of the ductile iron and carbon steel in the Hagerstown Inntn soil a; shown in Fig. -} reveal a similar rrlj. lion to that observed liclwecn die gnu iron and steel specimens in the earlier tests as shown in Fig. 2. This imlj. cates that with respect to weight Inn anti pitting, ductile iron corrodes in this soil at about the same rate as grai iron. Comparison of the corrosion data r,| ductile iron anti steel after exposure for 8 years (Fig. 3-7) shows that tinweight loss of the two materials urn about the same in three of the soib: Hagerstown loam, clay, and tidal marsh. In the other two soils. Sagemoor sandy loam and I-akewootl sand, the carbon steel corrotlcd at a liiglicr rate. With respect to pitting, the data show that the ductile iron and carbon steel corroded at about the same rale i r the moderately corrosive Lakewood sand and in the highly corrosive cbv and tidal marsh soils. In the letter two soils, specimens of both materials were perforated by corrosion within the 8-ycar exposure period. In the Hagerstown loam and Sagemoor sandt loam the maximum pit depths for the ductile iron were significantly greater than for the carbon steel. If'righl loss and iiio.ritmnn pit tIfflit arc plotted as a function of lime for ductile cast-iron (solid lines) and cnrbou-sicel (dashed lines) pipe buried tu the tidal marsh at Patuxent, Md. Summary In summary, the corrosion-time curves for specimens in these five soils show that with respect to weight loss, the ductile-iron specimens show the same general rates of corrosion for the 8-ycar exposure period in three of the soils as do the steel specimens. In the other two soils, ductile iron shows lower weight losses. The pil-depth-time curves show that in three soils the rale of corrosion of ductile iron is about the same as that of steel, and in the other two soils. CTD000156 Inel96! rirtt performance 65.1 the ductile iron corrodes at a greater tale. Tlic differences in weight loss or pilling shown by the specimens of ductile iron and carhnn steel may he iliown to lie significant in some of the soils by statistical analyses. Hut [ur practical purposes, as far as the corrosion engineer is concerned, the differences tire ol such small magnitude ilut the two materials can be con sidered, in general, to behave essen tially alike in the same environments. Hence, it can he concluded, after 8 years of exposure in the five soils where the specimens have been ex posed, that ductile iron and carbon steel corrode at nearly the same rate. Hccnuse one of die major conclu sions drawn on the basis of the Imigrangc studies from tin- earlier XIIS corrosion studies' was that lb- eommoiily-used ferrous metals, inelinling gray cast iron and curium steel, cor rode at nearly the same rate in the same soil environment, it also can lie concluded that gray iron and ductile iron corrode at nearly the same rate in the same soil environment. It is also obvious front the data in Table 2 and Fig. 3-7, that the materials corrode at different rates in different soils. Tit. 8. Specimens of Ductile Oast-Iron Pipe Exposed 8 Years at Five Test Sltea Tketop row shows pipe exposed in Sagemoor sandy loam (Site ./), Ilaycrslo-.en loam (Sue B), ami clay (Site C). The bottom row shows pipe exposed in I^akewood sand (Site I)) ami Tidal marsh (Site G). CTD000157 654 MELVIN ROMANOFF This is the same conclusion made in the 4-ycar progress rcjiort.- it lias been noted with reflect to the 8-year s|jcciiiieii.s, as was rqmrtcd in the 4-ycat rqiort, that the graphitic corrosion products of ductile iron had the same general characteristics and adhesive properties as were observed on specimens of gray iron in previous closures. Therefore, after 8 years of exposure to five different soils, it can be con cluded that ductile cast, iron and gnu cast in n corrode al nearly the s:uiV ran* in the same soil, and that th* | sit tern of corrosion and ih<- nature .i the cnrrosit.ii products are aU. similar in the same soil environments. Acknowledgments 'Die assistance of lhiijnmiti T Sanderson in various ca|ariiir> during the investigation is acknowledged. Acknowledgment is made also to Dr. * i__ j~l-.-f- ton ma i4ii jcii nut j*ai Tig. t>, Specimens of Carbon Bteel Pipe Exposed 8 Tears at Five Test Sites The top row shows pipe exposed in Sagentoor sandy loam (Site A), ItanerstKe* loam (Site /?), and clay (Site C). The bottom row shoivs pipe exposal m I niva'*1* sand (Site /)) end 7'idal mar.tlt (Site C). CTD000158 PirP. PERFORM AKCE *55 Frank L. LaQuc, International Nickel Co, for making (lie s|iecimeus of duclile-iron pi[>e available. :iikJ [o the fol -in organizations who provided test ..:rt and assistance during the Inina! aid removal of specimens: Bonneville Power Administration, US Dept, of Interior; Baltimore Bureau of Water vfiply; US Coast Guard Klertrouics Engineering Station, Wildwood, \'.J.; US Coast Guard Receiving Center, Cape May, N.J.; Patuxent Naval Air Station, l.exiiigtou I'ark, Mil. References 1. KoMANOFF, M. i/itJi'rornunJ Corrosion. .Vationul ISurcau of St.im|.tr<l\ Circular 57V, US Govt. I'ritilim' Oilier, Wash ington, D C. (1**57). 2. kiiMAN'Ol-P, M. Kxtcrior Corruvimi of Ca"t*Irun 1'ijK.*. Jour. I1JV 5<*: