Structural Considerations in the Design of the POLAR Class of Coast Guard Icebreakers Bruce H

Home , Icebreaker, Propeller

,,.W,+, THE SOCIETYOF NAvAL ARCHITECTSAND MARINE ENGINEERS 74 TrinityPlace,New York,N.Y.,10006 # *: P,PertobeDre$entedattheShipStructureSymPosiUm 8 :c W,sh,rwton,DC.,October68,1975 010 ~+ 2 Q%,,“,,*>>+. Structural Considerations in the Design of the POLAR Class of Coast Guard Icebreakers Bruce H. Barber, Luis M. Baez and Gary J. North, Members, U.S. Coast Guard Headquarters, Washington, DC. c Cowr,8ht1975 by The SOC’etY.f NavalArc17,tects,nd Msr,!,eEnE’neers

Artist’s conception of United States Ceast Guard icebreaker POW STAR (WAGB-10 )

structural failures were undertaken to ABSTRACT aesist in the selection of hull materials, Various finite element analyses The design of the Coast Guard’s were conducted for portions of existing POLAR Class icebreakers incorporated and proposed structural arrangements the results of four years of research, with special emphasis on the forward analysis and testing aimed at optimizing structure. With the coperation of other the structural design. Opration and agent ies, studies of the strength of sea load data were gathered by instrumenting ice led to updated ooncepts of loading existing icebreakers. Exten8ive tests that resulted in significant changes in and studies of available steels as we 11 design phi losophy snd refinements over as a methodical analysis of reported previous icebreaker designs.

c-1 —

INTRODUCTION The need has been apparent since the early 19607S for a new class of ice- Polar icebreakers onerate under the breaking ships to replace aging members most inhospitable ocean ~onditions the of the fleet and to undertake more ex- world has to offer. Designed for opti– tensive duties as Coast Guard Responsi- mum performance at low temperatures in bilities change and expand. The recent Ice fields varying from u“ifomn plate extension of the search for oil into the ice to deep windrowed ridges, they also Arctic region, for instance, appears transit areas of intense heat, high likely to require a strong Coast Guard winds, and extreme sea conditions. response capability in this area. In Their hull structure sees a variety of 1966 the Icebreaker Design Project was loads ranging from those thermally in- established in the Naval Engineering duced to concentrated ice Impact loads Division at Coast Guard Headquarters to to those impossible to design for, such initiate preliminary design work on a as grounding on uncharted rock pinnacles . new polar icebreaker. Tbe initial The icebreaker designer must allow for thrust of the effort was toward a nu- extreme loads on the one hand, but must clear–powered cutter, hut this was later pay attention, like all naval architects , modified to conventional diesel–electric to detail design, with a view toward and finally revised to include a gas- eliminating structu~al failures from turbine mode of operation. During the more subtle causes such as elastic in- existence of the Icebreaker Design Pro- stability, brittle fracture, and fatigue. ject, its personnel delved into the si&- nificant aspects of ship design as it The United States Coast Guard af- applied to icebreakers Its most signi- fords icebreaker support, both domestic ficant contributions to icebreaking and polar, to a variety of ship opera- technology were probably in the areas of tions conducted in areas made hazardous hull form. Dowerin!z predictions. and or impassable by ice. Typical polar hull stru; t~re, th; iatter including icebreaker missions include escort of material selection. In eXC’?SS of 100 vessels supplying outlying military or discrete projects were completed, many scientific stations: independent logis- dependent on Coast Guard-initiated re- tics SUDDOrt to Similar OUtDOSt S . and search. Literature searches “EIT con. ice and’~cean survey operations ;nd suP- ducted to insure that existing technol- port in both polar regions. ogy was not neglected. Project members received significant support from con– COaSt Guard icebreakers engaged in tractors, other government agencies, and these operations have recentlv consisted a large number of Coast Guard personnel. of six WIND Class vessels (Fikure 1) and the GLACIER (Figure 2) Budget reduc- By the time of the dissolution of tions and old age will have soon forced the Icebreaker Design Project in 1969, a the decommissioning of all WIND vessels basic preliminary design had been devel- exceut for WESTWIND and NORTHWIND. each oped. HUI1 fopm and principal dimen- of w~ich has undergone extensive ~truc- sions had been defined, rough arrange- tural and machinery renovation to prolong ments completed, a machinery plant size its service life. and type selected, a basic structural

Length over’all------269‘–or! Displacement to DWL--–--5,3OO long tons Length, DWL-–------25O ,-0” Displacement , maximum–--6,5l5 long tons Beam, maximum––- –––-63, -61, Complement ------l74 Beam, DWL––––––––---621 -07, Speed, knots, cruising––l6 Depth to main deck-- 37 T-9–l/2,! Propulsion–––––– –––––––– Die,sel-electric Draft to DWL-----–––25!–g11 SHP------10,OOO Draft , maximum---– ––2g, -l?< Number of screws ------2

Figure 1. WIND Class Profile and Characteristics + c-2 II“q r A 4 — w? ,.-. r: y-....> ,,- ~----- ,.-~ ..

Length overall------jlo t-7ts Displacement to DWL-----7,6OO long tons Length, DWL------Z9O, -O?l Displacement, maximum---8, 449 long tons Beam, maximum------74 9-3-1/21! Complement------lg7 Beam, DWL------72 q-611 Speed, knots, crui,si”g--l7.6 Depth to main deck-- 38v-4qt Propulsion------Diesel- electFic Draft to DWL------25, -g,, SHP------16,000 Draft, maximum------28* -611 NwnbeT of screws------2

Figure 2. GLACIER Profile and Characteristics

Length overall------OT1)-OT1 Displacement, maximum--- 13, l79 long tons Length, AWL------352,-O!1 Complement ------l38 Beam, maximum------83, -7rq Speed, knots, cruising--l7 Beam, DWL------–78 *-Ott Propulsion------Die.sel-electFic or Depth to main deck--49,-3–l/21, Gas turbine Draft to DWL------28, -011 SHP------18 ,000 D-E Draft, rnaximunl------31,-9!t 60,000 G-T EIisplacernentto DWL-11,000 long tons Number of screws–-–-----3

Figure 3. POLAR Class Profile and Characteristics

c-3 #--- arrangement chosen, and tentative scant- research twenty-five years of highly lings determined. In March of 1970, theoretical icebreaker and ice technol– these early beginnings were transferred OgY literature was obvious. Further, if to the existing Design Branch of the these studies “ei-e to be of any value Naval Engineer-ing Division for contract they had not only to be compared to the desizn development. Completion of the original design philosophy of the WINDS cont~act desi~n in 1971 ;as followed by and GLACIER but also comelated to O“T bid solicitation and award of a contract operational experience with these ships. to Lockheed Shipbuilding and Construction It was also necessary that the investi- Company, Seattle, to build one icebreak- gations put in proper perspective the er, the POLAR STAR (Figure 3) Delivery state of the art i“ countries such as of POLAR STAR took place during 1975. A Canada and Finland, which had been b“s- second ship, POLAR SEA, will be completed ily building icebreakers since the early in 1976. 1950’s.

STRUCTURAL DESIGN The design philosophy of the WIND Class took into account only a few major Structural studies within the Ice- ship parameters such as displacement, breaker Design Project began with an shell slope at the ice belt, ship length, assessment of existing ships GLACIER and beam, by use of the formula reported and WIND Class experience. total lin’a in Volume 54, 1946 of the SNAME Transac - some 220 ship-yeaks of ophration, w~s = [11 analyzed along with that of other ice- breake~s, domestic and foreign, to iden- tify structural inadequacies and to or- .—D— D d’+”z ganize this information in a format useful to the designer. Accumulated experience “reinvested” in subsequent designs is, of course, the essence of the clas- where P is the load per foot of water- SiC shiv desi.m Drocess. and the method line perimeter with the ship supported aPPears” tO re~ai~ its usefulness as tech- solely by ice at the waterline, D is the nology advances. displacement, m is the a“erage shell slope, and L is the waterline perimeter. The areas of structural concern The unit load P readily yields the load listed below were identified as requir- per frame which was in turn used to ing special investigation: design the ice belt plating and framing. The load per frame was envisioned as ● Load Definition quasi-concentrated at a pressu~e of 3,000 PSi and applied halfway between ● Framing System transverse frames in the desizn of the plating. The framing was the; designed ● Allowable Stresses to withstand the same load at midapan. The above design philosophy yields a ● Analysis Methods structure in which the framinK is incaD- able of “withstanding a unifor~ly distr~- ● Hull Material buted pressure of the order of the crush- ing pressure of ice and considerably ● Detail Design less than the distributed pressure that the plating can suppo?t. This “framing Although any ship design requires pressure capacity,, is less than 150 psi that these items be addressed, special for the WIND class and about 200 psi for emphasis was imposed in this case be- the GLACIER, both of which have expe~i - cause of (1) the lack of documentation enced damage in service. for many decisions reached in designing previous icebreakers, (2) obvious tech- Ice Strength nological advancements in the interven– ing years , (3) the accumulation of ex- There was therefore no question of pedience with post–war icebreakers and the need to establish valid design ice resultinz conceut realignments . and pressures based on realistic operating (4) the ~ealizaiion thai this bro,ject conditions. A thorough ice technology presented an opportunity to perform literature search indicated a wide dis- benchmark studies that could be refer- parity in the measur-ed physical proper- enced with confidence in future design ties of ice and, more significantly, work showed that most reports did not indicate important factors such as sample LOAD DEFINITION size and orientation, temperature, and salinity (Table I). Introduction This resulted in a decision to Determining rational design loads undertake an independent ice pressure for a ship of this size and type is a study Assistance was provided by the considerable undertaking. The need to Army ’s Cold Regions Research and Engi-

C-4 STRENGTH uE::cI{lPTIoN THICKNESS TEMP COMPRESSION TENSION SOURCE DEG F PSI PSI

Clear, cold, solid ice Not stated Not stated 3,000 250 H. F. Johnson SNAME-54 (1946)

Not stated Not stated Not stated 213 to 388 190 Fruhling –l°C of Koenigsberg (Markessa) (prior to 1900)

Clear fresh water ice Not stated Not stated 327 to 1,000 Not stated W. Ludlow pan (prior to 1900)

River, Kennebec Not stated -4°F 399 to 970 Not stated Prof. Kolster, Helsinki (prior to 1900)

Not stated Not stated 32°F 363 Not stated Pr’of. H. M. Mackay Prof. H. T. Barnes (1914)

St. Lawrence river ice Not stated 28°F 300 Not stated Ice of excellent quality 5-in. X 5-in 14°F 693 Not stated Prof. E. Brown Blocks loaded to 1,000 test pieces 2°F 811 Not stated (MCGiIl University) pounds in 2 sees

Hard old arctic ice Not stated Not stated 1,000 250 A. Watson, IME (1958)

Coast Guard criteria O°F surface based on temperature. 10-25 feet ?O”F bottom 600 pSi ..- Dayton [2] salinity, & ;trength profiles

Table 1. Ice Characteristics as re~orted i“ Volwne 67, 1957 of the SNAME Transactions [3] with Coast Guard data addea

:1. r’ neering Laboratory (CRREL) and the studies were documented by Dayton [21.

The following parameters were considered:

● Sample Size - Due to the meater incidence of faults ~id impurities, the average compressive strength decreases as the sample size increases.

● Sallnity - Varies through the thickness of sea ice; strength is In turn a function of salinity.

● Temperature - Varies through the thickness Strength decreases with increase in temperature.

From the above, strength profiles and average values for representative thicknesses of ice were calculated (Figure 4) . The following average desizn uressure values were then arrived at: c~nsidering data reliability and contact area factors for each condition o and in both cases departing from a maximum ice sample strength of 1,000 psi: Figure 4. Ice strength profiles, from [2] ● uniform static load for beset condition: Work of this nature included that by Nogid [4], White [5], Kheisin [61, Factors for and Tarshis [7]. The Icebreaker Design Project initiated additional studies, Sample size .4 resulting in contributions by Estes [8] Strength profile .75 and Dayton [2] . Contact area .5 Data reliability 2 Dayton applied computer techniques to Tarshis, p~ocedures to detei-mine im- Design pressure = pact load and area as a function of lo- 1,000 psi x .4 x .75 x .5 x 2 cation. The analysis was dependent on = 300 psi ship speed, hull shape, mass and rigid- ity of the ship structure relative to ● Uniform Impact design pressure the ice, and entrained mass of water. for bow and stern areas : A siemificant findin!z of this work was that-impact loads co~ld realistically be Factors for considered as uniformly distributed over .4 a large area, typically 48o square feet Sample size fo~ an 18,400-ton impact load. Strength profile 1 Contact area 1 Data reliability 1.5 Figure 5(a) illustrates the effect of location on total impact load as Desi En Dressure = determined by Dayton for one of several l,oolipii x .4 x lx 1 x 1.5 hull designs developed during the piw. = 600 pSi liminary design phase. It is apparent from the superimposed waterline that the greatest loads are predicted fop the Load Distribution shoulder area at about station 2. Fig- ure 5(b) provides a correlation with It was next necessary to determine WIND Class expeI.ience. The two-peaked the manner in which these loads were to curve indicates the relative nunbe~ of be applied to the hull. Since impacts reported damages “ersus longitudinal establish the maximum load to be absoFb - location, as compiled by D. E. Woodling ed by the shell strwcture, studies were and R. A. Yuhas of the Coast Guard. The carried out to determi”’e the distribu- higher peak OCCUFS at the bow screw bos_ tion of Impact loads on the hull under sing (the bow propeller was ne”eP in- varying conditions of speed, floe size, stalled on most ships of the class and ice thickness., and impact location. has been removed fmm the others ), which

* c-6 I-----Y— DEEP W, L. — LIGHT WL. DAMAGE LXS7/2/OUT/ON (25 YE+iSj b$’/A/oCLASS / i?O~ PROF/LE

D D WL Figure 6. POLAR Class icebelt cOnfigura– tion showing design pressures, from [9]

NEW Bo ~ $74 (b) CX?IGINAL * stresses were included along with ice 43$?/0 loading in the design of the ice belt. The decks and bottom structure were Figure 5. Bow impact locations , POLAR sized to avoid plate buckling failure Class (predicted) and WIND Class (actual) under wave-induced bending and hydro- static loads. The bottom plating thickness was further increased to account provides a “stop!t in heavy ramming con- for corrosion and to improve the resis- ditions and has pi-ecipitated nume~ous tance to grounding damage. structu~al ~aflu~es at station 1. (The faired stem of the POLAR Class and I.’eno- FRAMING SYSTEM vated WESTWIND and NORTHWIND IS exoected to alleviate this problem by redic;ig Various schemes of longitudinal shock loads resulting from sudden energy framing systems were considered and dis- dissipation during compI-essive loading carded because of weight considerations. of the ice by the steD. ) Davto”ts anal- Included among these was a system that ysis did not- include ihi bosiing area made use of extrusions incorporating up ~nmthus related better to the new hull to two frames and their corresponding As the figure indicates, WIND areas of shell plating. The study Class shell failures have also peaked in proved this to be heavier as well as the shoulder area whe?e the analysis costlier because of the unique pro- predicts peak loads. curement and fabrication problems Involved. Icebelt Definition The transverse framing system his- The Icebelt is that portion of the torically used for icebreaking ships was shell plating that is strengthened to therefore adopted. Supports for trans- accept the design ice loads. Figure 6 verse frames were readily available In indicates the extent of the icebelt on the form of decks; the considerable side POLAR Cla.SS ships. tankage pePmitted the use of decks or partial decks throughout the length and Longitudinal Bending depth of the hull. The framing was canted at the bow and stern so as to Studies were also made to confirm place the framing members more perpendi- the magnitude of wave-induced bending cular to the shell for more effective stresses as well as those experienced load-b earinz abilitv. The 16-inch frame when the bow is “beached” on an ice spacing agi~n” prove; to be the least- shelf. These studies confirmed previous cost alternative when compared to smal- investigations conducted with the MACK- ler grillages requiring more frames and INAW [101 in which it was concluded that lighter plating. the stresses induced by these loading conditions are of secondary importance Both WIND Class and GLACIER have, In the determination of the shell and within approximately the midships six- deck scant lings but nevertheless influ- tenths length, transverse frames suP- ence their design. Still-water bending ported by truss structure, Figure 7, I

Figure 7. Truss framing system of WIND Figure 8. Grlllage framing system Class and GLACIER of POLAR Class whereas other U. S. Coast Guard icebreakers (CACTUS. STORIS. etc. ) and mOSt EW- opean icebr; akeps eiploy deep webs and girders or grillage structure for shell support In all cases, however, bow and stern framing consists of grill age struc– ture. Detailed studies of each system, per fomned by the coast Guard and hy Lloyd? s Register of Shippin8 [11], cmn- pared the strength and weight of opti– mized truss and grillage structures. Particular attention was paid to the EifD w6e Wfsb Ewo behaviour of the structure under owP- (a) load conditions. As expected, the truss was shown to be the lesse~-weight system when compared on an equal design-strength basis It was the response of each to overloads that led to the selection of the grlllage, Figure 8. The truss system, in response to an overload condition, develops significant bending mo- ments at the connections which, combined with the axial loads on the compression members, could lead to a pro- gressive failu?e of adjacent frames through elastic instability, and subse- MD quent collapse of the shell framing. This mode of failure has been obsemred on present Coast GuaPd icebreakers. A stable grillage structure, on the other Figure 9. Alternate POLAR Class shell hand, may ‘experience local plastic de- framing details, (a) proposed and formation under overloads but will not (b) final necessarily lose its ability to sustain additional loads at the design load level. load would be applied to one transverse frame o“er it,? length between two decks Figure 9(a) illustrates a f~mnlng The purpose of the web frame-supported system that was under consideration into longitudinal member was to distribute the preliminary design phase of the the load onto adjacent frames. As the POLAR Class This system was consistent load criteria were refined, however, it with an early assumption concerning ilII- became appar-ent that this was not a pact load distribution, namely that the realistic load assumption and that the 600 psi impact load could be considered tion of plates and shells, is unrealis- distributed uniformly over a larger area tic if quantitative data for such mem– supported by many transverse frames. hers is required. This reduced the basis for selection to a simple one of minimum weight and cost, Fe” et al r81 studied the bow thus the system of Figure 9(b) was struct~re of an-i~ebreaker using SAMIS3, chosen. a matrix interpretive code which modeled the structui-e as a combination of beams, ALLOWABLE STRESSES bars, and shell elements. In order to save time and decrease the possibility Since the design pressure values of error, an input data generator was discussed ea~lier contained reliability developed to aid in providing descrip- faCtOFS , it was decided to design the tive data for nodal coordinates and shell structure to the yield strength boundary conditions The WIND Class when loaded amidships to 300 psi, the bow was modeled and it was determined uniform beset load, or at the ends to that relatively low pressures were suf- 600 psi, the uniform impact load. ficient to produce yielding; this was in agreement with p~evious observations . Plastic deformation will be expect- This study also reviewed the various im- ed when imposed loads exceed the design pact loading criteria and verified that loads Daytonts analysis predicts the the rise times of these loads were suf- formation of three plastic hinges at a ficiently gr’eat that dynamic effects 1,200 psi distributed load. Greensponvs could be disregarded. methods [12] indicate a permanent shell The detailed structural desire of plate set of O.09 inches at a uniform the POLAR Class utilized both manial and pressure of 1,150 psi. computer calculations. The program STRESS was used to analyze typical trans- ANALYSIS METHODS verse frames and the Neilsen Grillage Computer Code [18] sa” particular appli- The finite-element method of emal - cation in tbe design of side shell gril- ysis proved to be an effective tool for lages in way of tbe machinery spaces evaluating icebreaker hull structure. Genalis [133 in”estlgated the applicabi- HULL MATERIAL SELECTION lity of the Drofn’am STRESS1 to two- and thrie–dimensional frame analysis and Experience compared WIND Class transverse fPaMe S by the two methods. He concluded that the Shell structural failures experi- two-dimensional computer analysis was enced bv Dresent icebreakers. Drinci - effective but that the three-dimensional pally W?ND Class ships, have” a?ways been approach, which included the effect of cause for concern. Figure 10 is typical longitudinal members, was preferable, of major damages incurred by these ships. despite its greater cost, because of the On several occasions of simificant shell improved accuracy. He noted that the damage, the oDooTtunity wai taken to per- two methods ~redicted frame bendinc mo- form metallurgical examination on ments that w;re different by as mu;h as cropped-out hull plate [19, 20, 21, 22]. 33 percent. Lloyd$s Register of Ship- These analyses were invariably charac- ping also used STRESS for the GLACIER terized by comments such as, “eXtenSiV’? analysis in their GLACIER-MOSCOW com– brittle fracture”, “indicative of brit- parison in 1967 [11]. tle cleavage type failures”, or “highly notch–sensitive” In at least one in- The program FRANZ was used by stance it was noted that catastrophic I,loydTs for the MOSCOW analysis and by failure would have occurred had it not Kiesling [16], who studied the applica- been for the local nature of the load tion of available computer methods to and the lack of a significant tensile tbe structural analysis of an icebreak- field beyond the immediate damaged area. er. He cmnpa~ed STRESS and FRAN and concluded that the latter was more suit- Table II describes the physical able since it could (1) handle more com- characteristics of samples of plate plex problems, (2) more easily account from the WIND Class icebreaker USS ATKA for temperature effects, and (3) was (now USCGC SOUTHWIND) in 1965 [22]. Tbe nore efficient in the use of machine material is unnormalized Navy High-Ten- time. Kiesling cautioned that further sile Steel (HTS) conforming to [231, the development was necessary in structural HTS specification in effect in 1943. modeling techniques and in determining Refer; nce [19] reported on an analysis tbe loads acting on an icebreaker, and of plate samples removed from USCGC noted that reliance on these programs, EASTWIND in 1956, and concludes, “The with their two-dimensional represent a- combination of poor notch-toughness, low

- Structural Engineering Systems 3 Structw.al Analysis a“d Matrix Solver [14] Interpretive System [17] 2 Framed Structure Analysis Program [15] By fa~ the greater part of the Coast Guard’ s structural failure data has been generated by WIND Class ships. Although GLACIER has incurred damage, it is but one ship versus the WIND Class , six and it has a somewhat greater design vressure due to its ice frames beirm of HTS rather than mild steel as on WIiD Class hulls. The present author.? could locate no reports of analyses on GLACIER hull steel which is, like the WIND, S, HTS. Whether this steel was normalized is not mentioned in a“ailable documents, although the hull construction plans antedate by about one year the ilTS spec– ification that first ~equimd nornlaliz- ing.

It was appa~ent from WIN3 Class experience that an improved material should be utilized in any new icebreaker. The en”i~onmental conditions of low temD- erature and high impact loads, combined’ with a notch–sensitive material, promot- ed brittle fracture in extreme icebr’eak- ing conditions

Hull Steel Re~uiTements

Studies by Yuhas and Schumacher [25] of the Icebreaker Design PYoject iclenti– fied the following factors as significant in the selection of icebreaker hull Flgum 10. WIND Class shell steel: structure damage ● Low–temperature toughness The air temperatures experienced by polar icebreakers operating temperatures and impact is have been measured as low as considered responsible for brittle be– –500 F. Figuz-e 11 illustrates haviour of the plating dui_ing failure. the zone of transition from Nomnalizing of the plate material -50” F air temperature to +28° ii; the laboratory] resulted in a marked F sea water temperature. The improvement in notch- toughm?ss proper- chosen steel must be suffi– ties. It may be assumed that bad the ciently tough at these tempei-- plating been normalized (as now required atures to Dreclude brittle by specifications) , the extent of the fracture. However, as recent- casualty would ha”e been of a less ser– ly noted by Rolfe et al in ious nature. ” The HTS specification has Ships Structure Committee Re- included normalized plate since [24] port SSC-224 [26], the combi- was introduced in 1953. nation of toughness, stress

USS ATKA HTS PRESENT-DAY HTS [221 [24]

(a“erag,, of 4 samples)

Tensile Strength 81,000 psi 88,OOO max.

Yield Point 56,OOO psi 47,000 min.

Elongation, % 27 20

Charpy V-notch energy, ft-lb 14 to 26 100 @ +40° F !?+40” F (typical)

Table II. USS ATKA shell plate versus p~esent-day HTS

.,. c-lo level, and flaw size togethe~ determine the resistance of a structure to brittle fractui”e.

9 Fatigue Strength. As in any structural design, mate~lal selection and detail design go hand-in-hand in determining the resistance of the ice– TR4USITI0N breaker str-ucture to fatigue. — Although longitudinal bending 20NtE cycles are less than for most merchant vessels, Figure 12, and ice impact loads on a SEAW4TF2 @ + m“ F given area occur relatively )+ infrequently [2], the stress levels experienced by an icebreaker are often sewere. Re– Sistance to fatigue as a basic material characteristic as- sumes secondary importance Figurw 11. Minimwn temperatures acting when compaTed with the low- on DolaT icebreaker hull cycle fatigue strength of structural details. ● Yield Point and Ultimate ● Ease of Fabrication and Re- Strength Adequate tensile pair. Icebreaklng vessels properties are necessar-y to typically have heavy plate and insure that the structure will dense framing systems and withstand the basic loads while therefore require a relatively not imposing high wt!i~ht pen- large number of construction alties and fabrication diffi- man-hours, many of them spent culties. within areas of difficult ac_ cess. It is to the owner, s ● cost . Less is best, but since advantage to use a material material characteristics must that will not further increase be optimized, cost may be sub- fabrication difficulties due ordinate to other i-eq”irements. to more stringent forming or welding requirements

2s WN).IES07A , CANADA (*WAG=) x 20 4 C27,28291

15 UOOSd EP. STATE, kOLVEQlhlE3TATE @O YSARS) AS RE$WZTED IN VCLUME 73, 1965 OF THE S NA W E TRANSACTIONS 10 AND IN SHIPS STRUCTURE COMMITTEE RCPORT SSC - 164 [29, mJ

5 7=0LAE zclzeIEs,AKEIZ 30-Y12 LIFE 0 1 I 10 I02 K+ !04 10’ m’ 10’ Joe lo’ t4

N= NUb4bEU 0= TIMES THE INDICATED STSE5S WILL bE EKCEEDSD

Figure 12. Cumulative frequency distribution of polar icebreaker a“d merchant vessels, from [2] !--- C-n

— ● Corrosion Resistance. Since cantlv“u.meater. than that of A537. In all low-carbon steels tend to the final analysis , too little weight exhibit similar cor~osion DI-o- reduction and ioo much increased C;st perties in salt water, this plus the ready availability of ASTM- parameter will not generally A537 were the reasons for the elimina- influence the steel selection. tion of HY-80 as a hull material. Of greater concern are the corrosion Dronerties of the ASTM-A537 heat-a ffecked’ zone at the shell seams, where accelerated The steel described by ASTM Speci- corrosion has been noted on fication A537, adopted in 1965, is a WIND Class and GLACIER hulls. low-carbon pressure vessel steel avail– Insufficient data of this able in either a normalized (A577A) or tYPe was available in 1970 to ;l~:enched and tempered (A53iBj-coidi- aid in steel/filler metal Steel of this chemistry was, in selection, and the Coast Guard 1970 as now, seeing wide use in tbe ma- has since initiated a lo–year rine field, tYPiCal applications being program to study salt-”ater in carriers of liquified petroleum gas corrosion on welded samples (LPG) “all-hatch” ships, and offshore of Icebreaker plate. drilling and production platforms in temperate and arctic regions. It had Hull Steel Candidates been considered a leading candidate for uolar icebreaker hull steel since the The search for the hull steel that ;nception of the design project due to best met the above requirements led to its excellent low-temperature toughness, two prime candidates, HY-80 and ASTM- good tensile and fatigue properties, and A537 [Table 1111: relative ease of welding. Its material cost was about 1.Q time~ that of ABS HY-80 (MIL-s-16216) Class B as reported in SNAME T k R Bul- letin No. 2-11 [321. The advantages of HY-80 for icebreaker hull structure are obvious. Fracture mechanics methods indi– Cross-rolling provides it with essenti- cated that A537B plate had sufficient ally equal properties in all directions, Charpy V-notch energy in thicknesses ideal for impact loading normal to the through 1-1/2 inches to insure through- plate. Its low-temperature toughness is thickness yielding before fracture, in outstanding. Its yield strength of accordance with Rolfev s criteria of [33]. 80,000 psi minimwn appears to provide In addition, the method of [34] was used considerable potential fo~ weight reduc- to verify that a built-in su~face flaw tion. But HY-80 has clear disadvantages would not propagate to a dangerous size also. Cross-rolling? alloying, and ex– over a 30-year lifetime. tensive testing requirements increase its procurement cost to 3.5 times ABS Steel Selection and Application Class B (1964 data as reported in SNAME T k R Bulletin No. 2-11 [32]) Compared The tensile properties of ASTM (for now on an equal-thickness basis) A-537 are achieved by allowing carbon with A537, it is less attractive for content to a maximum of O .24%. The welding because: Coast Guard decided, however, that resol- ution of the toughness–tensile character- 9 A greater preheat/interpass istics trade-off should favor the former temperature is required. and investigated a carbon reduction to achieve this. The work of Roper and 9 Lower allowable heat input re- Stout published in Ships Structure Com- sults in lesser deposition mittee Report SSC-175 [35] and elsewhere rates. [36] indicated that a reduction to 0.17% would provide thick-plate (to 4 inches) ● A higher-strength, more costly A-577 material with towzhness far suDer– electrode is necessary. ior to that of ABS Clas; C while ten;ile strength was reduced to about 80 KSI, ● Joint preparation is more dif- and that underbead c~acking was elimi- ficult due to its greater hard. nated by using low-hydrogen electrodes . ness. During the polar icebreaker prelim- Not only do these factors tend to d~ive inary and contract design phases, J. W. cost upward, but they also increase the Kime and R. A. Yuhas of the Coast Guard difficulty of reDair work and make such met with major steel InanufactuTers to work in remote shipyards less desirable determine the optimum chemistry of an Furthermore, because of fatigue, buck– A537-type steel for this application. ling, and other considerations, full ad– They then developed a new specification vantage cannot be taken of the highe~ limitirm carbon to O.16% and increasing yield Strea.s to achieve weis!ht reduction. the Cha~py V-notch requirement to 20 f; i’ensile strem.th..U.... a.—.me..sure-of..— overload– lb @ .ho O F transverse (base metal & absorption capability, is not signifi - as-welded) compared with 15 ft lb @ ,-

* C-12 z H

!--- C-13 -75° F longitudinal for the ASTM material. This specification was titled llcG_A537M,, to indicate a modified A537 steel for Coast Guard icebreaker application. Steel company pricing methods were such that the identical steel would cost less if purchased under a new specification than as modified A537B.

Steel to specification CG-A537M saw its most essential application in the transition region (Figure 11) , where both extreme impact loading and low temperatures will occur. At higher 10- cations, low temperatures will be the dominant environmental hazard, compli- cated by unavoidable structural discon- tinuities; on the underbody, impact loading at water temperature will pre- vail. The use of CG-A537M was extended to include the basic hull envelope and other materials were incorporated as shown by Table IV.

Since icebreakers are historically NuM~Q OF CYC1 ES _ unable to retain bottom paint, a cOrrO- sion allowance was added to the icebelt Figure 13. Schematic fatigue design and all other underbody steel, including curves, from [2, 27, 371 the rudder. consisting of a l/b-inch addition b~yond that ~equired for strength. A current Coast Guard R 8 D details for any ship design, the method study is attempting to determine whether that it implies has not been rigorously present-day materials have potential as developed. The limitations are chiefly abrasive-resistant underbody coatings in the extrapolation of curve A, the for iceb~eaker application. Such a cumulative distribution of frequency, discovery would have significant impact to some curve B, the ship equivalent on the cost of future icebreakers. of the S-N curves C, D, and E. Such a POLAR Class ships will, when new, carry transformation involves altering the around some 175 tons of sacrificial fatigue load spectrum of curve A into a steel. representation of number of load applications versus constant mean stress DETAIL DESIGN ‘level. Although Nibbering [271 eXPlains one such method, he is careful to empha- AlthouEh data of the type presented size its approximate nature. schematically in Figure 13 [37, 271 would be most useful to the structural A further limitation on the use of engineer involved in the selection of this method appears to be the lack of

APPLICATION REQUIRED MATERIAL

Low-temperature Shell includin!z Icebelt CG-A537M Ice frames - CG-A577M Other shapes, Fabricated CG-A537M Rolled CG-A537M or ASTM-A537A or ASTM-A537B Flight Deck HY-80 Other Weather Decks (Main k 01) CG-A537M Internal Structure Adjacent to Shell or Weather Decks (plate) CG-A537M (rolle~ shapes) ASTM-A537A or ASTM-A537B Surrounding Large Deck Openings HY-80

Interior Structure Not Subject to Low Temperatures ASTM-A131

Superstructure ASTM-A131

Table IV. POLAR Class steel types

C-14 =. r..4,9,m’ IF1“.,.0’

I \ rN-2.4 “6”

FN-2 x ,03 I I — 4 ,x10’ ,x ,,9$ ,r10’ ,Xd

Figure 14. Relative fatigue resistance of structural details, ~rom [271

S-N curve data for specific Material/ detail combinations under loading conditions identical to those aboard ship.

The data Of [3B] (Figure 14) was useful to the extent that it provided a relative ordering of various joint designs in terms of fatigue resistance. Despite the low cycle predictions, the most fatigue-resistant details were selected for structure which would be expected to absorb impact loads, keep- Lng in mind the usual icebreaker hazards - extreme loads, low temperatures, the remote operating domain - and those nor- mally expected such as material flaws, construction-related structural notches, and corrosion. Figure 15 illustrates tYPical jOint designs provided on POLAR Class vessels. WAUNC1450 FQ.4M.s #AuffCL&O ,=%?A#E ~OOVC /CEbCLT 47 lC~MLT A tenet of sound structural design is to insure that instability failure does not prevent a member from absorbing Figure 15. Typical POLAR Class shell its design load. Lateral supports on framing details all heavily loaded framing are incorporated into the POLAR Class structure to preclude such failure. COMNENT

Additional detail considerations More extensive use of Icebreakers worth noting are that hull fr~ea are and ice-navigating ships is forecast for continuous through decks and that longi- the coming years. Various rule-making tudinal structure is continuous through bodies have responded with additional m. bulkheads. A general requirement, in- more severe structural requirements. fact , is that plate not be loaded The Canadian Arctic Shipping Pollution through its thickness in order to eliml- Prevention Regulations requi?e icebreak- nate the possibility of la.mellar tearing. er hulls capable of withstanding pl.es- Sures as great as 1,500 psi without ex- ceeding the yield stren~th of the hull structire. The Amt?I.ica~Bureau of Ship- ping has outllned requirements for new

C-15 L“

i 1 Ice Classes requiring shell design ice POLAR STAR costs some 90 tons and one- pressures of up to 234.5 psi for navi- quarter-million dollars. Assuming ade- gation in “extreme ice conditions!t. quate strength and ductile behaviour of The latter requirements alone are diffi. the material, how much plastic deforma- cult enough to attain. They are compa~- tion should be tolerated - structurally able to or exceed the actual strengths or even aesthetically - under extreme of many of today ts ‘!pola~,vieebmakeps. loads? The need for fw.ther resea~ch was never more appa~ent than now if these require- It appears that the de,sign of fFam- ments are intended to specify the true ing can likewise be improved. Figure 17 load-bearing capability of a ship $s hull. indicates the relative strengths of shell and framing for representative ships, As Figure 16 indicates, cum.ent data both existing and hypothetical, and com- on the world, s icebreakers continues to pares the load- cam.ying ability of plat- point to a correlation between ship dis- ing designed plastically, plating design- placement and plating thickness. Curves ed to yield, and framing designed to A and B describe older and more recent yield. The consistent differences be- trends respectively for arctic icebreak- tween shell and framing st~ength; “iewed ers, while curve C shows the present in the light of Coast Guard experience, trend for Baltic and Great Lakes ice_ aPPeaF tO require additional study. The breakers. Note that only curve C tends framing strengths shown assume that decks to indicate a limit on plate thickness. and bulkheads provide unyielding support for transve~se f~.mnes and lon~ltudinal Neither WIND Class ships nor GLACIER girders. This is an optimist~c assump- have sustained damage that can be att~i- tion, particularly for icebreakers where buted solely to plating failure. Defor- decks and bulkheads are not necessarily mation between frames is not apparent normal to the shell in the areas of even after thirty years of Arctic service greatest load. Furthermore, the framing accompanied by 15 to 25 percent plating strengths of Figure 17 do not account for deterioration. It therefore appears that the reduction in load-bearing capacity perhaps this plate has never been loaded caused by frame instability under heavy to the point that further increases in loads . This is due to shapes which are thickness will be justified. This issue inherently unstable because of section has immense impact when one considers asymmetry (such as inverted angles) or that one-eighth inch of shell plate on to frames which, even though symmetrical,

.50

1111 ]● WGk 76%41LE 3rEEL - & K&i. YIUD I I I I I I I . I / I 1 4 & 8 /0 /2 /4 /6 /8 20 22

DISPhICEMENT IN 70~s x lo Q o

Figure 16. Icebreaker shell plate thickness as a function of displacement

+_ -- c-16

/ 1 -.

m FRAMING STRENGTH AT YISLD

~ PLATING STRENGTH AT w ELD

= PLATING IN PLASTIC CONDITION i

Figure 17. Relative strengths of icebreaker bow shell structures

cannot be installed normal to the shell. The enlarged detail of Figure 18 illustrates the situation as found in varying degrees throughout the length of any ship, particularly those transversely framed, yet its effect on the ultimate strength of the plate-stiffener combination has yet to be assessed through the application of research. Structural fabrication cost must be \\ further addressed, including cOnsldeFa- ,* tion of weight, shipyard capabilities, and framing system options available to W-IELL the deslgnei-. Typical might be a study of framing systems such as that shown by ?Igure 19. This illustrates the concep- tualized framing of an icebreaker of about 300 feet owPall len~th, and sho”s BUD Longitudinally framed grillages in the ~ess contmmed areas of the hull, allow- Figure 18. Shell framing not normal to ing more extensive use of standard rolled the plate.ECombined beam and plate shapes. While somewhat heavie~, the de- section modulus does not represent the sign minimizes the costly custom fabrica- true ultimate strength in this situation. .

C-17 h–––– Figure 19. Conceptual longitudinal framing system with cant frames fore and aft tion of frames necessary with a trans- REFERENCES verse framing system. 1. H. F. Johnson, “Development of Ice- SUMMARY breaking Vessels for the U. s. Coast Guard”, Transactions , SNAME , This paper has only highlighted the Vol. 54, 1946. extensive research effort that attended the POLAR Class structural design. The 2. R. ~ , Dayton, ‘tPolar Icebreaker resulting CUtterS, although yet unproven, Preliminary Structural Design and are exoected to accomplish the Coast Special Studies”, Consultec, inc. Guardts missions with’ the utmost in Report 006, 1968 (AD 845-560) structural reliability. This work provides a foundation upon which future 3. J . G . GeI.man, ,,Design and Construc- desiKns. with new and additional require- tion of Icebreakers”, Transactions , ment;, !iill be based. SNAME, Vol. 67, 1959.

Changing design criteria are impos- 4. L. M. Nogid, ‘tImpact of Ships with ing new constraints on all ship designs . Ice” , (translation) , Transactions, Cost ceilings, once flexible, are now Leningrad Shipbuilding Institute, unyielding. Prevention of pollution is No. 26, 1959, P. 123. mandatory. More reliable performance is expected as new missions are added while 5. R. M. White, ‘tDynamically Developed manning levels are reduced. The consci- Force at the Bow of an Icebreaker”, entious designer can no longer rely on PhD Dissertation, MIT, 1965. intuition or on the extrapolation of past concepts to solve the new problems. He 6. D, ye. Kheysin, ‘vDetermination of must actively and objectively seek out Contact Forces when a Ship 1s Bow sound bases for his choices. The POLAR Strikes Ice”, Problemy Arktiki I Class design is but one example of his Antarktiki, Vol. 8, 1961, PP 67-74. dependence on the research establishment for the necessary answers to his inevit- 7. M. K. Tarshis, ‘tIce Loads Acting on able questions. Ships” (translation) , Rechnoi Transport, Vol. 16, No. 12, 1957. ACKNOWLEDGEMENTS 8. E. Fey, W. Pilkey, and P. Estes, The authors wish to extend their thanks ,,str”ctural Analysis of Polar Ice- to the following persons who were in- breaker Bows” , Illinois Institute volved in the POLAR Class structural of Technology Research Institute design and who kindly reviewed the draft, Project J6127, 1968. DrovldinE valuable comments and encour- agement: - Mr. R. B. Dayton, Jr. , CDR J. 9. L. C. Melberg Jr. , J. W. Lewis, W. Kime, CDR L. C. Melberg, Jr. , CAPT K. R. Y. Edwards Jr. , R. G. Taylor, B. Schumacher, and LCDR R. A. Yuhas . We and R. P. Voelker, “The Design of are further indebted to MI.. Kirk BuI.neP Polar Icebreakers”, Spring Meeting, for his review of the materials section, SNAME, 1970. to M?. Booker Smalley and Mr. William Law for their illustrations, and to Mm. Donna Bremer for her patient preparation of the manuscript.

4 Num~er~ in parentheses at the end Of I a reference refer to call numbers In the Defense Department Documentation 1 Center, Camerok Station, Alexandria, I Virginia, 22314. 1-

c-18 ~

10. J. Ormondvoyd and J. F. Shea, llAnal- tories Report No. 2941, 5 October ysis of the Hog-Sag Test Data of 1965. the MACKINAW”, University of Michi- gan Engineering Research Institute, 23. Ad Interim Specification 48S5 (INT) , August 1948. Bureau of Ships , 15 September 1941.

11. !,Large Polar Icebreaker for United 24. Military Specification MIL-S-16113B, States Coast Guard!r, Lloydrs Regis- 4 December 1953. ter of Shipping, R. T. A. Report No. 5048, 1967. 25. K. B. Schumacher and R. A. Yuhas, ,,Hu1lMaterial Sele Cti0r3 fOF Ice- 12. J. E. Greenspon, “An Approximation breaker Hulls” , ProJect No. 1739 to the Plastic Deformation of a (unpublished) , USCG Icebreaker De- Rectangular Plate Under Static Load sign Project, 1968. with Design Application!t, DTMB Report 940, 1955. 26. S. T. Rolfe, D. M. Rhea, and B. O. Kuzmanovic. “Fracture Control Guide- 13. P. Genalis, “Three-Dimensional lines for Welded Steel Ship Hulls”, Stresses in Icebreaker PrimaFy SSC Report No. 244, 1974. Structure stn,MIT Department of Naval Architectu~e and Marine Engi- 27. J. J. W. Nibbering, “Fatigue of ~eeying, Report No. 67-7, 1967. Ship Structures’!, Report No. 55S, Netherlands Research Center T. N. 14. STRESS , A Userrs Manual, The MIT U. of Shipbuilding and Navigation, Press. 1963.

15. FRAN : A User Ts Manual, IBM. 28. R. Bennett, ‘!Stress and Motion Measurements on Ships at Sea, Part 16. E. W. Kiesling, “Structural Analy- I and II” , Report No. 13, The sis Including Thermal Effects for Swedish Shipbuilding Research Assoc- Icebreakers’!, Southwest Research iation, 1958. Institute, 1967, (AD 680-070). 29. Idem, Part III, Report No. 15, 1959. 17. T. E. Lang, “Summary of the Func- tions and Capabilities of the 30. J. Vasta and P. M. Palermo, “An Structural Analysis and Matrix Engineering Approach to Low-Cycle Interpretive System Computer Pro- Fatigue in Ship Structures”, =- gram” , Jet Propulsion Laboratory, actions, SNAME, Vol. 73, 1965. California Institute of Technology, 1967. 31. D. J. Fritch, F. C. Bailey, and N. S. Wise, “Results from Full-Scale 18. R. Nielsen, Jr. , Analysis of Plane Measurements of Midship Bending and Space Grillages under Arbitrary Stresses on Two C4-S-B5 Dry Cargo Loading by use of the Laplace Trans- Ships Operating in North Atlantic formation, Danish Technical Press, Service” , SSC-164, 1964. 19b5. 32. “Guide for the Selection of High- 19. J. J. Gabriel and E. A. Imbembo, Strength and Alloy Steels” , T & R ,,Re~ear~h and ne”elopment On the Bulletin No. 2-11, SNAME, 1964. Notch-Toughness Properties of a Sample of Fractured Hull Plating 33. S. T. Rolfe, “Fatigue Criteria for Removed from USCGC EASTWIND (WAGB Structural Steels”, p~esented at 27q )”. New York Naval Shinvard the American Society for Metals Material Laboratory Proje~~ 5208-5, Educational Conferences on Fracture NS 011-084, 8 May 1956. Control for Metal Structures, Phila- delphia, Pennsylvania, January 1970 20. L. A. James, “Summary Report on an and Chicago, Illinois, May 1970. Investigation of the Fati!zue and Fractur; Properties of Se~ected 34, G. M. Sinclair and S. T. Rolfe, Hull Plate Samples from the Coast ,tA~~lytical Pyocedure for Relating Guard Cutter STATEN ISLAND” , Paci- Subcritical Crack Growth to Inspec- fic Northwest Laboratories, 14 Aug- tion Requirements”, U. S. Steel ust 1969. AFPlied Research Laboratory Project No. 39.018-007(26), NAVSHIPS Con- 21. T. J. Hogland and J. R. Kay, “Metal- tract Nobs-94535 (FBM). . lurgical Examination of Failed Steel Shell Plates from USCGC SOUTHWIND” . 35. R. D. Stout and C. R. Roper Jr. , Value Engineering Inc. Report 17-Rb- ,,Mechanical Properties of a High- 68, 15 August 1968. Manganese, Low-Carbon Steel fOr Welded Heavy-Section Ship Plate”, 22. USS ATKA (AGB 3) Fractured Hull SSC-175. Plate Analysis, Boston Naval Ship- yard Materials and Chemical Labora-

C-19 L–– 4“ 1 36. C. R. Roper and R. D. Stout, ‘fWeld- abllity and Notch-Toughness of Heat- Treated Carbon Steels”, Welding Journal Supplement, September 1968.

37. I. M. Yuille, “Longitudinal Strength Or ships”, RINA, 1962.

38. N. E. Jaeger and J. J. W. Nibbering, ,,Beam Knees and Other Bracketed Con- nections”, I. S. P., 1961.

I C-20