0MlR FFE COPY

US Army CorpsREPORT88u15of EngineersREPO T 8*15Cold Regions Research &

Engineering Laboratory

AD-A201 012Ship model testing in level iceAn overview

LECT E?

~lRBU'IN N T Aj~

Aploved for pulbUicrejecle~t~butoi~Unlimited

88 112.5 042

For conversion of S1 metric units to U.S.lBritishcustomary units of measurement consult ASTMStandard E380, Metric Practice Guide, publishedby the American Society for Testing and Materi-als, 1916 Race St., Philadelphia, Pa. 19103.

Cover: Views of ship model testing in theCRREL test tank.

I.

CRREL Report 88-15October 1988

*. Ship model testing in level iceAn overview

Jean-Claude P. Tatinclaux

~Accesiun or

NT•S ,CRA ,-DT;C TAB 0

Approved for public release; distribution is unlimited.

UNCLASSIFIED

SECURITf CLASSIFICATION OF THIS PAGE

Form ApprovedREPORT DOCUMENTATION PAGE oMB No 0704o-88Exp. Date lun 30. 1986

la REPORT SECURITY CLASSIFICATION lb. RESTRICTIVE MARKINGSUnclassified

2. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORT

2b. DECLASSIFICATION IOOWNGRADING SCHEDULE Approved for public release;distribution is unlimited.

4 PERFORMING ORGANIZATION REPORT NUMBER(S) S MONITORING ORGANIZATION REPORT NUMBER(S)

CRREL Report 88-15

PEO FORSJNG ORGANiZATION 6b. OFFICE SYMBOL 7a NAME OF MONITORING ORGANIZATION.s. &rmy (2O1dKegions Research (If applicable)and Engineering Laboratory CECRL

6c. ADDRESS (Oty, State, and ZIPCode) 7b. ADDRESS(City, State, and ZIP Code)Hanover, New Hampshire 03755-1290

Ba. NAME OF FUNDINGISPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION (if applicable) ILIR Program, DA Project 4A161101A91D/00/465

Sc. ADDRESS (City. State, and ZIP Code) 10. SOURCE OF FUNDING NUMBERSPROGRAM PROJECT TASK WORK UNITELEMENT NO NO. NO. ACCESSION NO

II. TITLE (Include Secunty Classification)

Ship M.odel Testing in Level Ice: An Overview12. PERSONAL AUTHOR(S)

Tatinclaux, Jean-Claude P.13a. TYPE OF REPORT 13b. TIME COVERED 14 DATE OF REPORT (Year, Month, Day) 15 PAGE COUNT

I FROM TO October 1988 3916. SUPPLEMENTARY NOTATION

17 COSATI CODES 18. SUBJECT TERMS (Continue on reverse of necessary and identify by block number)FIELD GROUP SUB-GROUP Model Ice Ship models

Model tests Towing tank tests

19. ABSTRACT (Continue on reverse if necessary and identify by block number)

This report presents a general discussion of model testing of ships in level Ice. The main points coveredare: '1) modeling criteria for ships in ice, which must take into account the presence of a solid boundaryat the water surface;t2) types of model ice used In various tanks--saline ice, urea-doped ice, EG/AD/Sice and synthetic ice; 3) techniques for growing model ice sheets, and achieving and monitoring the re-quired ice properties; 4) limitations of both model ice and property measurement techniques;,5) modeltesting procedures for EHP and SHP tests and their limitations; 6) comparison between model test resultsand available full-scale trials data; 7) existing empirical and analytical or semi-analytical algorithms forpredicting ship performance in level ice; 8) current research at CRREL and other research facilities toImprove modeling techniques and data interpretation; and 9) novel bow designs for ice-transiting vessels.

20 DISTRIBUTION i AVAILABILITY OF ABSTRACT 21 ABSTRACT SECURITY CLASSIFICATIONI UNCLASSIFIEDUNLIMITED 0 SAME AS RPT 0 OTIC USERS Unclassified

22a NAME OF RESPONSIBLE NOIVIDUAL 22b TELEPHONE (include Area Code) 22c OFFICE SYMBOLJean-Calude Tatinclaux 603-646-4100 CECRL-EI

00 FORM 1473, 84 MAR 83 APR edition may be used until exhausted SEC-RITY CLASSIFICATION OF THIS PAGEAll other editions are obsolete UNCLASSIFIED

...... -. ~ l ! I I ! no ,, u m , l , , . .

PREFACE

This report was prepared by Dr. Jean-Claude P. Tatinclaux, Research Hy-draulic Engineer, of the Ice Engineering Research Branch, Experimental Engi-neering Division, U.S. Army Cold Regions Research and Engineering Labora-tory. It originated from a presentation given at the September 1986 meeting of theNew England section of the Society of Naval Architects and Marine Engineers.The CRREL research mentioned in the report was funded under several con-tracts with the U.S. Coast Guard and under the In-house Laboratory IndependentResearch (ILIR) program, DA Project 4A161101A91D/00/465.

This report was technically reviewed by Dr. George D. Ashton of CRREL, andby Dr. Stephen Jones of the Institute for Marine Dynamics, National ResearchCouncil, St. John's, Newfoundland. Additional comments and suggestions wereoffered by David Baker of Melville Shipping Ltd., Ottawa, Ontario, and by Dr.Gary Timco, Hydraulics Laboratory, National Research Council, Ottawa, Onta-rio.

The contents of this report are not to be used for advertising or promotional pur-poses. Citation of brand name does not constitute an official endorsement or ap-proval of the use of such commercial products.

ii

CONTENTSPage

A bstract ................................................................................ iP reface ................................................................................. i iN om enclature ......................................................................... vIntroduction ........................................................................... 1General considerations .............................................................. 3Ice properties ........................................................................... 5

M odulus, E .......................................................................... 5Characteristic length, I ............................................................ 6Flexural strength, af ............................................................... 6Com pressive strength, or .......................................................... 6Shear strength , a .................................................................... 7Poisson's ratio, v ................................................................... 7Fracture toughness, Kic ............................................................ 7D ensity, pi ........................................................................... 7Ice-hull friction factor, fi ......................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

M odel ice ............................................................................... 8Synthetic ice ........................................................................ 8Colum nar saline ice ............................................................... 8Colum nar carbam ide ice .......................................................... 9Fine-grained ice ................................................................... 10EG/AD/S m odel ice ................................................................ 10

M odel test procedures ................................................................ 11Ice growth and monitoring ...................................................... 11E H P tests ............................................................................ 12SH P test ............................................................................. 13

Test data analysis--comparison with full scale ................................. 14Analysis of test results ........................................................... 14Comparison with full-scale data ................................................ 16

Analytical and empirical predictors .............................................. 16Em pirical predictors .............................................................. 16Analytical and semi-analytical schemes ...................................... 18

Current research efforts in ice modeling ......................................... 19International cooperative research .............................................. 19Ice testing ........................................................................... 2DCRREL research on ship-ice interaction ....................................... 2

Novel icebreaking bow designs ..................................................... 23Conclusions and final remarks .................................................... 26Literature cited ........................................................................ 25

iii

.ILUSTRATIONS

Figure1. Measurement of ice-hull friction coefficient .............................. 112. Example of resistance signal in level ice .................................. 133. Example of nondimensional resistance test results ....................... 134. Example of propulsion tests results (captive model) ....................... 155. Full-scale resistance prediction from model test results ................. 166. Comparison between predictions and full-scale trials

m easurem ents ............................................................. 177. USCG Katmai Bay performance curve- in level ice ....................... 198. Floe size distribution from tests in level ice with an idealized

icebreaker bow .............................................................. 219. Dimensionless test results .................................................... 22

10. Effect of floe size on propeller performance ............................... 2211. Effect of floe size and velocity on propeller torque standard

deviation ................................................................... 2312. Expected effect of ice density on ice floe trajectory along a ship

hull .......................................................................... 2313. Original Waas bow concept ................................................... 2414. Recent developments in Thyssen-Waas bow .............................. 2415. Wartsila experimental bow .................................................. 25

TABLES

Table1. Existing ice testing facilities ................................................. 12. Variables governing ship-ice interaction .................................. 4

iv

NOMEENCLATURE

A mean plan area of ice floes Pw shaft power

B ship beam Q propeller torque

C, Cauchy number R total resistance in ice

d propeller diameter Si Pi/P

D ship draft SQ torque standard deviation

E ice modulus t thrust deduction factor

Fd densimetric Froude number Th propeller thrust

Fn Froude number V ship speed

A ice-hull friction coefficient W i width of ice floe

g gravity P3 ship block coefficient

hi ice thickness Y water specific weight (pg)

J advance coefficient (V/nd) A ship displacement

In Atkins "ice" number X geometric scale

kj, k 2 coefficients X, velocity scale

K4 torque coefficient 77P propulsion efficiency

KT thrust coefficient water dynamic viscosity

Kic ice fracture roughness v ice Poisson's ratio

Ic ice characteristic length p water density

L ship length p ice density

Le length of ice floe ac ice compressive strength

n propeller rotational speed Of ice flexural strength

p porosity Ys ice shear strength

V

Ship Model Testing in Level Ice: An Overview

JEAN-CLAUDE P. TATINCLAUX

INTRODUCTION from 1955 to 1975, only three ice towing tankswere built, one in Finland, one in West Ger-

Model testing in ice of ships and other many, the other in the United States. In thestructures is a relatively recent branch of following decade (1975-1985), ten such tanksmarine engineering, with the first refriger- were built in Japan, Northern Europe (in-ated towing tank being built in 1955 at the cluding the USSR), and North America.Arctic and Antarctic Research Institute of the Table 1 recapitulates the ice tanks currentlySoviet Union in Leningrad. However, with operating, with their major characteristics.the rapid development of the arctic and sub- All types of ships with varying missionsarctic regions of the globe, spurred by the dis- and, thus, requirements are encountered incovery of large mineral and petroleum depos- the Arctic and Antarctic, the Northern Atlan-its, the fields of arctic engineering, in tic, the Baltic sea, and as far south as thegeneral, and ice engineering, in particular, Great Lakes and St. Lawrence seaway; theyhave rapidly grown in the last decade or so. range from tankers (oil and liquified natu-For economic and geopolitical reasons, the ral gas), to cargo and container ships, supplynumber of ships transiting ice-infested wa- ships, research vessels, and national securi-ters has greatly increased. Correspondingly, ty or Coast Guard vessels.better and new ship designs specifically As in conventional towing tank tests, theadapted to the requirements of a particularly main purposes of ship-model resistance andharsh environment led to the development propulsion tests in ice tanks are: 1) to predictand construction of new facilities devoted to ship performance in ice for a given design, 2)ship model testing in ice. In the two decades to suggest modifications to a proposed design,

Table 1. Existing ice testing facilities. The basin length given may includethe length of the trim tank.

Main dimensions Year ofL x B x D (m) opti alson Mudcl ice

CANADA

NRC/IMD 94.0 x 12.0 x 3.0 1985 EG/ADInstitute for Marine DynamicsSt. John's, Newfoundland

NRC/Hydraulics Laboratory 18.0 x 7.0 x 1.2 1980 EG/AD/SOttawa, Ontario

Arctec Canada Ltd. 30.5 x 7.3 x 1.4 1981 SyntheticCalgary, Alberta

Arctec Canada Ltd. 30.5 x 4.9 x 1.5 1977 SalineKanata, Ontario

Table 1 (cont'd). Existing ice testing facilities. The basin length given mayinclude the length of the trim tank.

Main dimensions Year ofL x B x D (m) operation Model ice

FINLAND

Wartsila Arctic 77.3 x 6.5 x 2.3 1983 FG-salineResearch CenterHelsinki

GERMANY

HSVA/Hamburgische 78.0 x 10.0 x 2.5/5.0 1984 Urea-dopedSchiffbau Versuch. 30.0 x 6.0 x 1.2 1971 Saline/UreaHamburg

JAPAN

SRI/Ship Research Institute 35.0 x 6.0 x 2.1 1981 SalineTokyo

NKK Tsu Research Laboratory 20.0 x 6.0 x 1.8 1982 Urea-dopedTsu City, Mie

IHI/Ishikawajima- 7.0 x 3.4 x 0.9 1982 Urea-dopedHarima Heavy IndustriesYokohama

MSB/Mitsui Ship 5.0 x 1.5 x 0.8 1985 Urea-dopedBuilding CompanyIchihara City

MHI/Mitsubishi 20.0 x 9.0 x 2.3 1986 Urea-dopedHeavy IndustriesNagasaki City

USA

ARCTEC Offshore Corporation 25.6 x 10.7 x 2.4 1982 SyntheticColumbia, Maryland 30.5 x 3.7 x 1.8 1974 Saline

Cold Regions Research and 36.5 x 9.1 x 2.4 1978 Urea-dopedEngineering Laboratory (CRREL)Hanover, New Hampshire

Iowa Institute of 20.0 x 5.0 x 1.2 1980 Urea-dopedHydraulic ResearchIowa City, Iowa

USSR

Krylov Research Institute 30.0 x 6.0 x 1.8 1986 Saline

Arctic & Antarctic 13.5 x 1.9 x 1.3 1955 SalineResearch InstituteLeningrad

2

and 3) to investigate new concepts and de- namely thickness, density and mechanicalsigns. As a loose kaalogy with ice-free (clear properties, and last but not least the ice-hullwater) tests eli!.er in calm water or in waves, friction coefficient. All of these parametersmodel tests "ri ice can be divided into two cate- must be scaled to satisfy similitude. There-gories, namely tests in level ice and tests in fore, ship modeling in ice involves hydrody-ridges. At this time procedures and tech- namics as well as mechanics of materials toniques for tests in level ice are fairly well es- account for ice deformation and rupture. It istablished and documented in the literature thus strongly dependent on the progress made(e.g., Schwarz 1983; Free, in press), while in the field of ice mechanics. However, themodel tests in ridges are far more recent and term "ice" encompasses many different ma-their procedures are still being developed. terials: freshwater ice, first-year sea ice andModel tests in level ice have also been the pri- multi-year sea ice, to name a few, with vary-mary concern of the Committee on Perfor- ing crystalline structure and correspondingmance in Ice-Covered Waters of the Interna- behavior under load. In spite of the difficul-tional Towing Tank Conference (ITTC), ties outlined above, the art of modeling ship-which has only recently begun to tackle the ice interaction has made significant advanc-difficulties involved in model testing of ships es in the past decade to the point that resultsin ridges. The present paper addresses only from model tests can greatly aid the navalmodel testing in level ice, where there still architect in the design of an ice-transitingremains some controversy, as discussed in ship.the main body of the paper. Perfect physical modeling of any phenom-

The views expressed in this paper are those enon requires exact geometric, kinematiccurrently held by this writer. It is likely not and dynamic similitude. Exact similitude isonly that they will change with time as more ensured when all independent, dimension-experience is gained and progress is made, less parameters involved take the same val-but also that some of them will be challenged ues at model scale as at full scale. For theby other researchers in this exciting field of particular problem of ship-ice interaction,ice model testing. these dimensionless parameters are listed in

Table 2, where the selected nondimensional-izing variables are ice thickness (hi), water

GENERAL CONSIDERATIONS density (p) and gravity (g).Since the gravity, g, cannot usually be al-

As in ice-free model tests, the dependent tered in the model tests, and the water in thevariables to be measured in ice model tests tank has nearly the same properties (densityare the total resistance (R) for EHP (effective and viscosity) as sea water, it is well knownhorsepower) tests, or, in propulsion or SHP that identity of the Froude number and the(shaft horsepower) tests, the propeller thrust Reynolds number between model and proto-(Th) and propeller torque (Q), from which oth- type cannot be achieved. Since tests in ice areer ship propulsion characteristics such as usually made at relatively low velocities,shaft horsepower, propulsion efficiency and corresponding to a maximum of 6 to 8 knots (3thruct deduction coefficient can be calculat- to 4 m/s) at full scale, the resistance compo-ed. nent due to the water itself is usually a small

The difference between model tests in ice- percentage of the total resistance and eitherfree waters and in level ice is the presence of can be neglected altogether, or accommodateda solid, flexible, floating boundary at the free with a minor correction of the test results. Insurface. This adds considerably to the diffi- this respect, it should be noted that the pres-culty of the tests by greatly increasing the ence of turbulence stimulators at the bow, anumber of independent variables or parame- common practice in open water tank tests, isters that need to be scaled. These variables, not acceptable in ice tests since even a smalllisted in Table 2, include not only the usual protrusion in the model hull will unrealisti-ones, namely ship geometric characteristics, cally increase the ice resistance. The mainship speed, propeller speed, water properties concern of the engineer and scientist in(density and viscosity) and acceleration of charge of ice tests is, therefore, to model asgravity, but also the ice characteristics, correctly as possible the ice properties accord-

ing to the modeling law selected.

3

Table 2. Variables governing ship-ice interaction.

Dimensional DimensionlessType of variables variables parameter

Ship Length L L/hicharacteristics Beam B B/hi

Draft D D/hiPropeller diameter d d/hi

Speed V F,= V/ gh or

F d =V/-(A p/p)gh'

Propeller speed n J= V/ndGravity g 1

Water Density p 1properties Viscsity p p VL

Thickness h1Density P P i /PPoisson's ratio V UModulus E E/p gh, or

E/ar orCn= pWIE

Ice Characteristic length 1, Ic / h,properties Flexural strength o" f arP pghi

Comp. strength a, crc/a fShear strength as cr8 /a r

Practure toughness Kl/= p 2

K Pj Vh l TK -

1.5or KIcpighi

Ice-hull friction factor fA fA

NOTES:

The ice characteristic length, 1c, is defined by:

1/4

lB 12 (1-v 2

F is the Froude number based on ice thickness.Fd is the densimetric Froude number.J is the advance coefficient.C. is the Cauchy number.I,, is the "Ice" number propose, by Atkins (1975).

4

.. . II I I I I a I I -I I

The current practice in ice modeling is to and Technology, and others. The proceed-

follow the Froude scaling law. If the geomet- ings of specialty conferences such as the Iceric scale is selected first, X = LA, = proto- Symposia of the International Association for

type length/model length, the ratio of proto- Hydraulic Research (IAHR), Port and Ocean

type velocity to model velocity is given by Engineering under Arctic Conditions(POAC) conferences, and Offshore Mechan-

X = VVm = XO.5. ics and Arctic Engineering (OMAE) confer-ences are an important source of up to date in-

Conversely, it can be envisioned that the formation.available towing capabilities will limit the Even a cursory survey of the literaturerange of model speeds and therefore dictate shows that ice mechanical properties are

the velocity scale XV. The geometric scale of highly variable. The values measured de-

the model would then be selected as 2 = . pend upon the type of ice (i.e., freshwater iceOther factors that also affect the choice of the or saline ice) and its structural characteris-model geometric scale are: model tank size, tics (i.e., grain size, crystallography and po-model cost, minimum model ice strength for rosity or brine volume in the case of sea ice),which test results are reliable, characteristics temperature, rate of strain, and even on theof available dynamometers for measuring measurement techniques. To address thisresistance, thrust and torque, or even size of latter problem, the ITTC Committee on Per-available model propellers. The model scale formance in Ice-covered Waters and thefinally selected is a compromise between the IAHR Committee on Ice Problems are at-often conflicting above considerations. tempting to standardize the experimental

Once the geometric scale has been selected, techniques for measuring ice properties boththe scales of all the other test variables and in the field and in the laboratory (Schwarzthus their range of values to be achieved in the 1979). It should be mentioned at the outset thatmodel tests are specified by the requirements measurement techniques do vary betweenthat the dimensionless parameters in Table 2 field and laboratory conditions (Timcotake the same values in the model as in the 1981b), and that the values attributed to theprototype. This implies in particular that ice natural and even more so to the model icethickness, ice modulus, and ice strength be should be treated more as indices than as ac-scaled by the geometric scale X, while its frac- tual mechanical properties of ice as a materi-ture toughness should be scaled by X1 .5. The al.parameters v (ice Poisson's ratio) and fi(hull-ice friction coefficient), which have no Modulus,Edimensions, should have the same values in The ice modulus, E, used in engineering isthe model as in the prototype. The first re- actually the initial tangent or effective modu-quirement is to know the range of prototype lus determined from the load-displacementvalues for the ice properties that must be mod- curve of a plate, a cantilever beam or a sim-eled. A brief discussion on ice properties is ply supported beam. The rate of load applica-given in the following section. tion should be high enough for creep deforma-

tion of the ice to be negligible or correctionsneed to be introduced when interpreting the

ICE PROPERTIES test data. This effective modulus is known tovary with ice salinity and temperature, i.e.,

Much research has been devoted to the field with porosity. Vaudrey (1977) gives the fol-and laboratory measurements of ice mechan- lowing relationship between E and porosity,ical properties. Relevr'nt information can be p, of sea or saline ice:found in Weeks and Assur (1967), Lavrov(1969), Schwarz and Weeks (1977), Tryde E (GPa) = 5.25- 13 fp(p _ 0.1).(1979), Mellor (1983), in the chapters on icephysics and ice mechanics of books devoted to For freshwater ice (p=O), the effective mod-ice engineering (Michel 1978, Ashton 1986), ulus predicted by the above equation is 5.25and in specialized scientific journals such as GPa, which is within the measured range of 4Journal of Glaciology, Cold Regions Science to 6 GPa. The modulus can be as low as 1 GPa

5

for warm sea ice with a porosity of 0.1 or strength of columnar S2 sea ice in terms ofgreater (brine volume of 100 0/ooo or more). the ice porosity, p:Other reseachers claim a linear variation ofE with p. af= 960 (1 - 2 p) kPa

The ice Young's or elastic modulus ismeasured usually by high frequency vibra-tional methods, i.e., by measuring the rate of valid for p < 0.11. For p > 0.11, rf remains ap-propagation of small-amplitude, high- proximately constant, equal to 250-350 kPa.frequency waves or pulses in the ice. The val- For columnar freshwater ice, it was foundues obtained in this manner are nearly twice that the flexural strength obtained from beamthose of the effective modulus. tests (cantilever or simply supported beams)

depended on the direction (upward or down-Characteristiclengthle ward) of load application (Gow et al. 1978).

When analyzing plate or beam test results The values of arf obtained with the ice top fiberfor the determination of E, an often implicit in tension is 1.25 to 2 times that measuredcalculation is that of the characteristic with the bottom fiber in tension. This differ-length, 1c. This length is a measure of the size ence is attributed to the anisotropic crystal-of the deformation zone of the ice plate or line structure of columnar ice, where thebeam when subjected to a vertical static load. grain size increases from top to bottom. ForThe two quantities E and lc are related by sea ice, however, there is no apparent differ-

ence in strength between the two loading di-3 4 rections (Tabata et al. 1967, Kayo et al. 1983).

E ] Also, the flexural strength of freshwater ice12(- V2) depends on the measurement technique, sim-ply supported beam tests yielding values asmuch as twice those obtained from cantilever

for a plate, and beam tests (Gow et al. 1978, Timco and Fre-derking 1982, Timco 1985). No such effect of

3 11/4 measurement technique on flexural strengthh was observed for sea ice.

Pg In view of the above, it is worth emphasiz-ing again that af should be considered as a

for a beam. flexural strength index and that, whenTo determine E from /c, one has to take the presenting data on af, one should also indi-

ice thickness, hi, into account. Also, it can be cate the measurement methods and condi-seen that an error of 18% in Ic leads to a 100% tions.error in E. Therefore, scaling the character-istic length may be preferable to scaling E. In Compressive strength, ameasuring l c by the plate method (Sodhi et al. Since the maximum ice forces exerted on a1982) for both lake ice and first-year sea ice, structure occur when ice fails in crushing,Sodhi et al. (1985) found that the ratio 1/h i for the compressive strength is the property thatlake ice was on the order 20, while for sea ice has received the most attention from re-it was in the range of 10 to 15. searchers. Through careful testing, it has

been determined that this property depends onFlexural strength, af the following parameters: ice type, grain

From results of cantilever beam tests and size, temperature or porosity (or both), strainsimply supported beam tests, it was found that rate and test conditions (i.e., uniaxial test vsoa decreased from approximately 1 GPa for triaxial test, relative orientation of force ap-freshwater ice to about 300 kPa for sea ice with plication to crystalline structure, etc.) (Tim-a porosity of 0.2 or greater. Thus the ratio of co and Frederking 1986). For the purpose ofeffective modulus to flexural strength, E/af, the present paper, i.e., for modeling of shipsvaries approximately between 5000 (for fresh- in ice where the loading rate is high, it maywater ice and cold first-year sea ice) to 2000 be sufficient to say that ac in the horizontal di-(for warm, weak sea ice). Vaudrey (1977), in rection is 3 to 5 times af, while in the verticalagreement with Weeks and Assur (1967), pro- direction it can be as much as 10 times af.posed the following relationship for flexural

6

Shear strength, or boratory ice is affected by the presence of im-Data on ice shear strength are very few pri- purities such as air bubbles or brine pockets.

marily because no fully reliable methodology The density of bubbly freshwater ice can be asto measure this ice property has yet been de- low as 890 kg/m3. For sea ice, the usual rangeveloped. On the basis of limited field experi- of density is from about 890 kg/m3 for cold,ments, Frederking and Timco (1986) have low salinity ice to 930 kg/m 3 for ice with highproposed the following equation relating the brine content.vertical shear strength of columnar ice to iceporosity, p: Ice-hull friction factor, fA

When a ship penetrates into ice, a signifi-an(kPa) = 1500 [1 - 1.60 1 cant component of "tie resistance to motion

may come from e friction of the ice on thehull. This fricti, .aal force is assumed to be

which is in agreement with the commonly ac- proportional to the force exerted by the ice nor-cepted assumption that or, is 1.5 to 2 times the mal to the ship-hull, and the coefficient ofice flexural strength. proportionality is the friction coefficient fA.

Numerous studies (e.g., Ryvlin 1973, Cala-Poisson's ratio, v brese et al. 1980, Oksanen 1983, Saeki et al.

To quote Mellor (1983, p. 53) "the value of v 1984, Akkok et al. 1986, Tatinclaux et al.for non-saline ice of very low porosity is 1986) have been conducted to determine the ef-about 0.33 ± 0.03, and variation with porosity fect of various parameters on fi. It was foundin sea ice is likely to be within the limits of that ice hardness, relative velocity betweenuncertainty for the pure ice value." It is com- ice and hull, wetting properties of the hull, -mon to take a value of v = 0.3 for sea ice and etc., could affect fi, in addition to hull rough-model doped ice (either saline or urea ice). ness characteristics such as roughness aver-

age and morphology of the roughness ele-Fracture toughness, Kji ments. Here again, the methodology and ap-

Until very recently, fracture toughness paratus used to measure fi may affect the val-had been a property totally ignored in the ues obtained. Most studies were made in theanalysis of ship-ice interaction. It is, howev- laboratory or under controlled conditions.er, receiving increasing attention and its The few direct measurements of fi on realimportance in ship modeling in ice is being ships (Hoffman 1985, Liukkonen 1985, Enk-investigated (Jones 1986, Parsons et al. 1986). vist and Mustamaki 1986) gave inconclusiveIt characterizes a material's resistance to and even inconsistent results. Low frictioncrack growth. It is likely, therefore, to control coatings, the best known of which is Inerta-the size of the resulting pieces when the mate- 160, have been and are being developed. Inrial fails in fracture (fragmentation). De- practice, it is generally assumed that a newfects or flaws in the material will decrease its ship hull has a friction coefficient of aboutfracture toughness. Air bubbles in freshwater 0.1.ice and brine pockets in sea or doped ice are Incomplete as it may be, the above discus-such defects. This material property is possi- sion should give a fair idea of the difficultybly the most difficult to measure experimen- in determining the ice properties. Values fortally. It is therefore not surprising that the these properties should always be accompa-data published in the literature for freshwater nied by a careful description of the test meth-and sea ice exhibit considerable scatter (see ods and conditions used to obtain them, asMellor [19831 for references). The data avail- well as information on ice structure, salini-able in the literature show values for the frac- ty, etc. It is worth repeating that, in fact, theture toughness that vary for freshwater ice values measured do not actually represent thefrom 60 to 300 kN m - 3 / 2 , and for sea ice from material properties of the particular ice test-30 to 140 kN m-a/2 . ed, but should be considered as indices to be

used for modeling or for comparison betweenDensity, pi model test results.

While the density of pure ice has long beenestablished at 917 kg/m3 , that of field and ]a-

7

MODELICE 60 mm, with most ice sheets (33 out of 40 or83%) between 15 and 35 mm thick. The

As discussed previously, once the geomet- flexural strength ranged from as low as 10ric scale of the model has been selected, the kPa to about 80 kPa; however, 34 out of 40 icemechanical properties of the model ice ought sheets (85%) had a flexural strength betweento be in a specified ratio with those of the field 10 and 20 kPa. The range of E/af cited is 700 toice to be modeled. There are currently five 5000 with an average of 1450 and with 90% oftypes of model ice used in the various tanks, the data points lying between 700 and 2000.as indicated in Table 1, namely synthetic ice, The ratio of compressive strength to flexuralsaline columnar ice, columnar carbamide- strength, ac/af, ranged from 0.6 to 2.3 withdoped (urea) ice, fine-grained ice and EG/ an average of 1.14, which is low compared toAD/S ice. Except for the first, all types of mod- that of sea or freshwater ice. Finally, the ice-el ice are grown from water baths containing model hull friction factor, fi, was given asone or more chemicals or dopants. These do- ranging from 0.07 to 0.28, which covers thepants are partially trapped between the ice entire realistic full-scale range that can begrains, thereby reducing the strength of the expected. Schultz and Free did not give val-ice. Further adjustment in the strength char- ues for the density, the shear strength or theacteristics of the doped ice can be made some- fracture toughness of the synthetic model ice.what by varying the growth temperature and They did, however, mention that "the size ofprimarily by raising the ice sheet tempera- the broken ice pieces resulting from ship orture to within a few degrees of the melting structure interaction with the synthetic ice ...temperature for a certain amount of time appears to compare very favorably with the(warmup period) prior to the tests. An excel- corresponding piece size observed in fieldlent discussion of the morphological proper- trials of icebreaking vessels and structuresties of ice grown from a doped solution was (p. 239)." They also cite one definite advan-given by Timco (1979). This section discuss- tage of the synthetic ice over doped model ice:es the main properties of these types of model it is handled at room temperature and itsice with their advantages and disadvantages properties remain unchanged for days.as can be gathered from the open literature.

Columnar saline iceSynthetic ice The first dopant used to create model ice in

The more widely known synthetic ice a refrigerated ice tank was sodium chloride.(MOD ice) is a wax-based mixture prepared By varying the salt concentration in the wa-in a liquid form and poured on the surface of ter bath and consequently the brine contenta traditional towing tank where it is allowed within the ice, almost any required reductionto set. The exact composition of the mixture is in flexural strength can be achieved atadjusted to achieve as closely as possible the growth temperature, which is normally wellrequired density and mechanical properties. below 0°C. This is still the technique used inSuch a synthetic ice was used as early as 1970 the ice tanks of the Arctic and Antarctic Re-(Crago et a]. 1970). Its rights were acquired by search Institute in Leningrad, U.S.S.R.Arctec Inc. (now Arctec Offshore Corpora- However, high salinity model ice has an ex-tion) of Columbia, Maryland, where it has tremely low strain modulus E, and the ratiobeen further refined. It is a combination of a E/af is usually no greater than 500. This ice"secret ingredient" and flat, round plastic is therefore excessively plastic, and breaksbeads or miniature disks, 3 to 4 mm across. into floes that are much larger relative toSince the exact composition of this synthetic those observed in the field. Schwarz (1975,ice is proprietary, no other ice tank besides 1977) suggested use of a low salinity icethose of Arctec Offshore Corporation in the grown from a water bath with a salt concen-U.S.A. and of Arctec Canada, Ltd., has any tration of 1.6 %c or less. The method suggest-experience with this model ice and there has ed involves three steps:been no independent study of its properties.Schultz and Free (1984) presented the results 1. When the water bath has nearly reachedof measuring the properties of 40 model ice its freezing temperature, a fine mist issheets. The ice thicknesses ranged from 10 to sprayed into the cold air above the tank. The

8

L

mist freezes in fine crystals that settle on the at CRREL yielded values of 0.93 g/cm 3. Awater surface to initiate the ice sheet. This significant advantage of carbamide over saltwet-seeding technique results in a columnar is that it is not corrosive. It is also a very com-ice sheet with small crystals, one that closely mon, fairly inexpensive chemical used, inresembles the crystalline structure of sea ice. particular, as a fertilizer.

Many ice tanks - Arctec Offshore Corpo-2. The ice is grown at as low a temperature ration, Arctec Canada, CRREL, HSVA, NKK

as possible (on the order of -200C), primarily - have adopted carbamide ice as model ice.to reduce the growth time to a minimum, but Hirayama (1983), in an extensive study ofalso to minimize the thickness of the nearly carbamide ice sheets grown from a 0.95%freshwater top layer. urea solution in water, confirmed most of

Timco's findings. However, in the experi-3. Once the desired ice thickness is nearly ence of this writer, while values of Eaf of 2000

reached, the temperature in the ice tank is are possible for relatively thick and strongraised to about 00C. During this warmup peri- ice, values of 1000 to 1500 are more common.od, the brine pockets in the ice melt, thereby For thin, weak ice, i.e., for ice sheets lessincreasing the liquid volume within the ice than 3 cm in thickness and with a flexuraland decreasing the flexural strength. With strength less than 30 kPa, the ratio E/af maythis procedure, Schwarz showed that values of even drop below 1000. Correspondingly, whilethe ratio Eof close to 2000 were possible as the ratio l/h i is on the order of 12 for ice sheetslong as the ice strength was above 40 to 50 4 cm thick or more and with a flexuralkPa. These were confirmed in an indepen- strength of 40 kPa or above, it may becomedent study by Timco (1981a). However, when less than 10 for thin, weak ice (i.e., 2 cm thickfurther reduction in strength was required, and with ar = 20 kPa). For these reasons, thisi.e., for a geometric scale greater than 15 to writer discourages users of the CRREL ice10, the ratio E/af could become as low as 500. tank from conducting tests at thicknesses ofAnother disadvantage of salt as a dopant is less than 2 cm or ice flexural strengths lessthat it is highly corrosive. Its use leads to high than 20 kPa. In this writer's experience, themaintenance costs for the steel components of results of such tests would be unreliable.the ice tank, towing carriage, and other Another disadvantage of urea-doped ice isequipment. its two-layer structure: the top, incubation

layer can be 3 to 10 mm in thickness and isColumnarcarbamide ice significantly stronger than the bottom, den-

After an extensive study of possible do- dritic layer. For that reason, the flexuralpants, Timco (1979, 1980, 1981a) suggested strength measured with the load applied up-carbamide (or urea) as an alternative to sodi- ward (bottom surface in tension) is only aboutum chloride for high modeling scale (>20). 40% of that measured with the load appliedFrom test results made on 4- to 5-cm thick ice downward (top surface in tension). Thesheets grown from a 1.3% by weight urea so- thickness of the top layer can be minimizedlution in water, Timco concluded that carba- by growing the ice at the lowest room tempera-mide was superior to low-salinity saline ice ture possible and ensuring that the water tern-and that scale factors up to 40 were possible perature at seeding is nearly equal to itswith this new doped ice. Over the range of freeezing point.flexural strength from 15 to 70 kPa, the ratio Finally, Timco (1985) measured values ofE/af remained larger than 2000, the compres- the critical stress intensity factor Kic from ansive strength in the vertical direction varied average of about 6 kPa m1/2 at ar = 40 kPa, tobetween 110 and 250 kPa and that in the hori- an average of 17 kPa m1/2 at af 130 kPa.zontal direction varied between 110 and 160 When extrapolated to full scale according tokPa. The shear strength under vertical load- flexural strength ratio, these values of KI, areing increased from 30 kPa at af = 15 kPa to 70 significantly greater than the range of 30 tokPa at e1f = 70 kPa, while that under horizon- 100 kPa m1/2 for sea ice with a porosity of 0.13tal loading increased from 35 to 65 kPa. The and that of 90 to 150 kPa m1/2 for ice with a po-density of carbamide ice was found to be on rosity of about 0.02 (Urabe and Yoshitakethe order of 0.95 g/cm 3 . Measurements made 1984, Timco and Frederking 1983). It can be

9

conjectured that the hard top layer of carba- tions. Also, the rate of growth of this ice is 5 tomide ice may be at least partially responsible 7 mm/hour compared to 2 to 3 mm/hour for thefor its high critical stress intensity factor. other types of model ice, which permits aAnother reason might be that model ice has much faster turn-around time in testing.fewer flaws than real ice. This high fracture Finally, the density and friction factor oftoughness of carbamide ice, even with a high the FG-ice are said to be comparable to that ofratio of E/af, may.be one reason that the size of columnar sea ice or model ice. No measure-the ice floes produced with this model ice re- ments of its shear strength or critical stressmain relatively larger than those observed intensity factor were reported.during full-scale trials. No independent testing of this fine-

While in the opinion of this writer carba- grained model ice confirming the above re-mide ice was a definite improvement over sa- sults is known to this writer. A similar fine-line ice, it is not yet the ideal material for grained ice is now in use at the Nippon Ko-modeling ice-structure interactions. kan K.K. ice tank, where carbamide rather

than salt is the dopant. The growth procedureFine-grained ice of FG-ice at NKK is also somewhat different

Almost concurrently with the development from that developed by Wartsila's research-of carbamide ice at the National Research ers, but results in a comparable model ice.Council in Ottawa, Enkvist and co-workersfrom Wartsila Arctic Research Center EG/AD/S model ice(WARC) in Finland were developing anoth- Timco (1986) introduced a new type of mod-er model ice (Enkvist and Makinen 1984). It el ice for use in refrigerated ice tanks. Theis a saline, fine-grained ice similar in struc- ice is grown from an aqueous solution ofture to snow ice (ice type T-1) grown from a three dopants, ethylene glycol (EG), aliphatic2% NaCI solution. Its structure is nearly ho- (very low sudsing) detergent (AD) and sugarmogeneous, as opposed to the columnar struc- (S). According to Timco, the ethylene glycolture of sea ice and carbamide ice. It is grown serves as a low molecular weight dopant thatby continuously spraying a mist of tank wa- is trapped in the ice as impurity pockets to re-ter above thL water surface. Since the air tem- duce the strength of the ice. The aliphatic de-perature is well below 0°C, the mist freezes in tergent reduces the surface tension of the so-small crystals that deposit on the water sur- lution and allows more of the ethylene glycolface. This procedure requires that the spray to be trapped by the ice. The resulting icenozzles be mounted on a carriage that contin- should be nearly single-layered, as opposed touously travels back and forth along the ice saline or carbamide ice. Finally, the sugartank. Careful control of the carriage speed, inhibits the lateral growth of the ice plateletsnozzle arrangement and discharge is neces- to produce a fine-grained ice.sary to create an ice sheet of uniform thick- The methodology to grow and temper thisness. Newly formed ice sheets have usually new ice to achieve prescribed thickness andtoo low a strength, and additional freezing mechanical properties is basically the samewithout spraying may be necessary. Accord- as for carbamide ice. Tests were carried outing to Enkvist and Makinen, the ratio E/ar, (E by Timco (1986) on 4-cm-thick ice sheetsmeasured by the plate deflection method) is grown from a solution with concentrations ofon the order of 1200 ± 200 for ice thicknesses of the three chemicals of 0.46/0.032/0.049% re-20 to 40 mm and flexural strengths from 13 to spectively. He reported that a flexural28 kPa. According to its developers, the uni- strength from 20 to 100 kPa could be obtained,axial compressive strength of this model ice and that the strength measured by upwardis somewhat low. loading was about 70% of that measured by

According to Enkvist and Makinen, the downward loading of the cantilever beam.main advantage of this ice, dubbed FG-ice, is The ratio of unaxial compressive strengththat it is more brittle than either saline or car- with horizontal loading to flexural strengthbamide ice, and therefore breaks into small- (aclaf) was about 3, while with vertical load-er floes than either. The resulting icebreak- ing this ratio was on the order of 5. The ratioing pattern and flow of ice pieces along a of strain modulus E (plate deflection method)ship's hull are similar to full-scale observa- to at was found to fall within the range 1500-

10

2500, similar to carbamide ice of same thick- tests in ice sheets whose thickness and flexu-ness. Finally, the KI, of the EG/AD/S ice ap- ral strength are scaled by the geometric scalepears to be much lower than that of carbamide X. The other properties are then only meas-ice and better simulates the critical stress in- ured and recorded. In all cases it is impor-tensity factor of sea ice over the usual range tant not only to report the measured values ofof scaling factors. these properties but also their measurement

From the above results, it appears that EG/ techniques if meaningful comparisonsAD/S ice is a significant improvement over among test data are to be ettempted.urea ice as a modeling material. So far it has Flexural strength is usually measuredbeen adopted only by the Institute for Marine from in situ tests of small cantilever beamsDynamics in St. John's, Newfoundland. Be- and the effective or strain modulus, E, by thecause of its cost (about twice that of urea ice) plate method. The kinetic friction coefficientand the time involved in fine-tuning the is measured by pulling a loaded ice samplegrowth and testing procedures for each partic- over a flat area of the ship hull or a test boardular ice tank, other organizations, which had whose surface was prepared in the samerecently adopted carbamide ice, are awaiting manner as the ship hull (see Fig. 1). Theor conducting independent studies of this new pulling force is measured and divided by themodel ice before deciding whether to adopt it total normal load to yield the friction coeffi-or not. It can be mentioned that the St. John's cient. Tests may be conducted at differenttank initially experienced severe growth of pulling speeds and normal pressures to deter-bacteria. This bacteria problem was resolved mine possible effects of these two parametersby deleting the sugar from the dopants and on fi (ITTC 1984). Test methods for other prop-installing a sand-type filtering system for erties may vary greatly from one test facilitythe tank water. to another when those properties are meas-

ured at all. For refrigerated facilities, thesize of the tank and its refrigeration charac-

MODEL TEST PROCEDURES teristics will dictate the time for ice to grow toa specified thickness and the tempering time

Ice growth and monitoring required to achieve a particular flexuralOver the usual full range of geometric strength. When the target flexural strength is

scale, X = 10 to 40, no existing model ice can nearly reached, there is usually only a nar-scale all the properties of sea or freshwater row time window during which the ice proper-ice. It is then customary to conduct ship model ties must be measured and the tests per-

ice -

we'(ghts °e " 1

4---

- Loa d Testam Cell S urface

a. Diagram of friction test apparatus.

Figure 1. Measurement of ice-hull friction coefficient.

11

A70

b. Friction test apparatus in operation.

Figure 1 (cont'd). Measurement of ice-hull friction cofficient.

formed. This lack of available time often analog signals be done at high enough ratesprecludes measuring all ice properties other to avoid aliasing effects. Once the towingthan af and E, which are considered as the carriage has reached a steady speed, the mod-most important to measure, unless sufficient el resistance is averaged over a sufficientpersonnel are available to conduct the meas- travel distance (one model length or more) tourements simultaneously or nearly so. To yield reliable values. Since model tests in iceadd to this problem, the weaker the ice, the are far more expensive than those in ice-freenarrower the time window available and the water, as many tests as possible are to be con-more difficult and time-consuming the ducted in one ice sheet.measurements. In addition to resistance, the most common

quantities measured during EHP tests areEHP tests heave, pitch and roll.

Resistance or EHP (effective horsepower) The tests results are sometimes presentedtests in level ice are conducted in a manner as the total resistance, R, versus velocity, V.similar to traditional resistance tests in However, the whole range of tests usually re-calm water. The model hull is connected to a quires more than one ice sheet, and the icetowing carriage, free to heave, pitch, roll and characteristics - thickness, strength, strainpossibly yaw. Depending on whether the tow- modulus, etc. - will vary slightly from oneing mechanism is stiff or soft, the hull may ice sheet to the next. This writer prefers to plotbe totally or only partially restricted in surge the data in dimensionless form as shown inand sway. One major difference between Figure 3.open water and ice resistance tests is that in In addition to quantitative measurements,the latter the hull motion is only steady in the qualitative test documentation by means ofmean. As the hull penetrates into and crushes photograph, films or video recording are ex-and bends the ice, the resistance increases tremely useful. In particular, underwateruntil the ice fails, at which point it decreases photography or video showing the motion ofsuddenly. This quasi-periodic or cyclic phe- the ice floes along the mv'4e l 'hull can be jsednomenon, an example of which is shown in to indicate potential interference between iceFigure 2, requires that digital sampling of the and appendages. These visual aids may sug-

12

0.6

600-f~

400-

200- Level I1ce

0 5 10 15 20 25 30 35Time (s)

Figure 2. Example of resistance signal in level ice.

80

4-

2 - Sheet Io 2

0 0.2 0.4 0.6 0.8 1.0Fn,

Figure 3. Example of nondimensional resistancetest results.

gest modifications to hull design and append- and the test data are interpolated to zero pull toage location to minimize such interaction, determine the self-propulsion point (rpm,

thrust, torque) at the particular speed. In theSHPtests free-model tests, the propeller speed is usual-

As in open water tests, propulsion or SHP ly set at a particular value, and the corre-(shaft horsepower) tests in ice can be conduct- sponding model velocity is then measured,ed either with a towed model or with a free together with propeller torque and thrust.model that is tethered to the towing carriage Since model speed in ice is only quasi-only by the power and analog data signal steady, this method requires that the maincables. In the former method, two to three tests carriage be able to follow the model within aat different propeller speeds for each model fairly close distance, and model speed or po-speed must be conducted. The force (pull) be- sition relative to the main carriage, or both,tween carriage and model must be measured, must be measured. This test method has the

13

advantage of yielding the self-propulsion In all cases, because of the scatter inherentpoints without any interpolation or other ad- in test data from ice model tests, some sort ofjustments of the test data. On the other hand, regression analysis is used to smooth the re-the method requires a rather long tank, espe- sults. The analysis may be made on the mod-cially at high speed, to allow the model to el data or the extrapolated values, in dimen-reach its quasi-steady state, and is, therefore, sional or dimensionless form. Based on thismore appropriate to the longer ice basins, author's experience, and until a more reli-

Since one of the goals of propulsion tests in able theoretical analysis of ship-ice interac-ice is to investigate potential ice-propeller tion is available, the form of the regressioninteraction, high-speed data acquisition sys- should be kept as simple as possible. It shouldtems are required to sample the thrust and be selected by visual inspection of plots of thetorque signals at high enough rates (about 10 dependent variables versus the independenttimes per revolution), variables that are believed to be of primary

Examples of propulsion tests results under importance. Since any regression analysisthe captive-model method are shown in Fig- is valid only over the range of test parametersure 4. As for resistance tests, visual docu- (or the corresponding full-scale values), thementation is very useful, especially of any final results should not be extrapolated be-ice-propeller interaction. yond that range, which needs to be specified.

In the case of resistance tests, the dataanalysis method used by this author consists

TEST DATA ANALYSIS - first of subtracting the open water resistanceCOMPARISON WITH FULL SCALE from the total resistance in level ice to obtain

the net ice resistance. The open water resis-Analysis of test results tance is measured by towing the model in the

Ship model tests are conducted to predict ei- ice model basin free of ice. The dimension-ther the power required to achieve design per- less net ice resistance is plotted versus theformance or the performance to be expected Froude number with the dimensionless icefor a given shaft power. In the first case, pre- strength as a parameter. A regression analy-dictions are usually plotted as power versus sis is performed on the basis of these plots, thespeed in level ice for several ice thicknesses. result of which is applied to full-scale condi-In the latter case, the expected speed at full tions to predict the ice resistance of the proto-power is plotted versus ice thickness. type. The open water resistance of the proto-

Each ice testing facility has its particular type, determined i'rom traditional open watertechnique for analyzing, interpreting and tests, is added to the ice resistance to give theextrapolating the test data to full-scale condi- total resistance in level ice at full scale.tions. One technique is to perform a regres- While the open water resistance is usuallysion analysis on the raw test data. For exam- small compared to the ice resistance, in thinple, the measured ice resistance is assumed to ice at high speed (6 to 8 knots [3 to 4 m/s]) itbe a polynomial function of speed, ice thick- can become a sizable component of the totalness, and ice flexural strength. Another resistance. The results of EHP and SHPmethod is to break the total resistance into model tests are used to evaluate the thrust de-several components and perform regression duction factor, t, and propulsive efficiency,analyses, assuming different forms of the re- q , as functions of the advance coefficient J =gression equation for each component. Still V/nd. Until better understanding of ice-another technique is to make similar analy- propeller interaction, and possible corre-ses in a dimensionless form. sponding scale effects, is reached, the model

Extrapolation of the model test data to full- results for t(J) and ?7p(J) are assumed to be ap-scale conditions is made either strictly by us- plicable to full scale.ing the Froude scaling law or by also intro- Examples of full-scale performance pre-ducing empirical correction factors based on dictions following the above technique of datapast experience or on available full-scale analysis are shown in Figure 5.data (see, for example, Enkvist and Musta-maki 1986).

14

218

0.4-

o 8.2-

30-1a . 4

a10 20 30 40 50 613 70 SEC

NRC TEST 0005

a. Initial test results.

Thrust Torque(N) (N }200- , 400 -J- P

Test Series

- Pull

Thrust . 6

10 Torque 6, a 350 -12

pul 10 0

(N) / 'O

10

-59

• 4 • 00

I l 1i50 1 71 50 39 6 7 8 9 X 05

n2

(rDM2)

b. Determination of self-propulsion test point.

Figure 4. Example of propulsion tests results (captive model).

15

TMrum Tor-ue

1001

60-

(lb/ton)

HULL

Wed-10D

20 .Con-100Wed-fit

- Rus- fIfI , I I i

0 2 4 6 8V (knots)

Figure 5. Full-scale resistance prediction frommodel test results.

Comparison with full-scale data with full-scale measurements for both ofThe validity of model test results can be as- these ships are shown in Figure 6. As can be

certained only by comparison with full-scale seen, the power predictions from model teststrial measurements. There are very few ice- are consistently greater than field measure-breakers and other ice-going vessels that ments. This author believes that this discre-have been tested and whose full-scale trials pancy is at least in part attributable to incor-results are not proprietary. Even among the rect modeling of the ice piece size (Keinonenfull-scale data available (e.g., Vance 1980, 1983) resulting from incorrect strain modu-Vance et al. 1981, Edwards et al. 1981, Mi- lus, fracture toughness and density of thechailidis and Murdey 1981) there are often model ice. This may lead to exaggerated, un-gaps in the data presented, or the full-scale realistic ice-propeller interaction in theconditions are significantly different from model.the scaled up model test values. This is not al-together surprising when one considers thedifficulties involved in full-scale ice trials. ANALYTICAL AND EMPIRICALFurthermore, it is extremely rare to encoun- PREDICTORSter in the field level ice sheets devoid of snow,while the model ice sheets are always so. In In parallel with the development of labora-addition, the presence of snow may signifi- tory model testing techniques, efforts havecantly increase the full-scale resistance, de- been made to develop predictor equations orpending upon the thickness and consistency numerical schemes for the level ice resis-(dry or wet snow) of the snow cover. tance and propulsion performance of ice-

In this author's opinion, some of the best going vessels. These equations or schemessets of full-scale data currently available are can be divided into two categories: 1) empiri-those collected by Vance (1980) and Vance et cal and 2) analytical or semi-analytical.al. (1981) with the USCGC Katmai Bay, andby Michailidis and Murdey (1981) with the EmpiricalpredictorsCCGC Franklin. Comparison of the ship per- Empirical predictor equations for the re-formance predictions from model tests con- sistance or propulsion characteristics in lev-ducted at CRREL (Tatinclaux 1984a.b, 1985) el ice, or both, have been proposed by Kashtel-

16

hl1 a. (in)

0 6 12 18 2414 1-7

- \k12 -6

20C-=600ePo200 10 - -5

.. v

.. --- - V 8-4 V.

' - (knots) l (m/s)-6- 800 ontDesign

6 - Point 3

- - 4- 2

2-- - Full-scole K \ I

0 020 0Predicted Res,stance lWN) him0 . Im)

a. USCGC Katmai Bay.

5"0

0MeosudoelM&,

Predicted Resitanr (Mw) IO M

0 5 0 1

0 ~~

Poer , r *-/

0 -/7 "J25 -"a / /-

-500'- 7 / .- -5

- ./ V.-7

/- -2 /2

0 500 000Meo eoire

Thru~t tkN)

b. CCGC Franklin.

Figure 6, Comparison between predictions andfull-scale trials measurements.

17

yan et al. (1968), Enkvist (1972), Lewis et al. Once this limiting ice thickness is esti-(1982), and Tsoy (1983). The equations were mated, and when the maximum ship speed inobtained by multiple regression analysis of open water is known, the ship performanceavailable full-scale and model test data. As curve, maximum speed vs ice thickness, canfor any empirical formula, these equations be drawn under the assumption that it is line.are valid only over the range of parameters ar. Such performance curve was estimatedfor which they were determined. In particu- for the USCGC Katmai Bay and comparedlar, the data bases for all the equations were with model tests predictions in Figure 7.obtained from classical icebreaker hullshapes. Therefore, it is unlikely that the pre- Analytical and semi-analytical schemesdictors may be used for the new hull shapes All analytical and semi-analytical deri.currently under development and discussed vations of ship resistance in level ice (MiIain a following section. no 1973, Carter 1982, Kotras et al. 1983) avail.

The set of equations proposed by Tsoy able in the literature assume that the ice(1983) was found by this writer to be of particu- resistance can be divided into several compo-lar practical interest. It predicts the limiting nents, namely a breaking resistance due toice thickness capable of being broken by a ice crushing, shearing, bending or bucklinggiven ship, i.e., her design point in level ice. (or all four), a frictional resistance, and aGiven the total shaft power, P, in kilowatts, component due to ice floe submergence andand the propeller diameter, d in meters, the entrainment. The various authors follow ei-thrust (Th) developed at near-mooring operat- ther a force balance (Carter, Kotras et al.) oing conditions is estimated by an energy balance (Milano) approach. The

mathematical models of Milano and CarterTh = k1 (4p 0 Pw) 2/3 (kN) appear to be purely analytical, since the only

inputs required are the ice properties and thewhere k1 is an empirical coefficient: k1 = 0.78 ship hull geometry. The resistance equationfor a single-screw ship, k1 = 0.98 for a double- proposed by Kotras et al. is considered semi-screw ship, and k1 = 1.12 for a triple-screw analytical, since it still contains empiricalship, assuming all propellers to be identical coefficients that need to be adjusted for eachand that power is equally distributed among ship studied from a few model test data orthe propellers. full-scale trial data; the equation may then be

The limiting ice thickness, h i in meters, applied to conditions other than those used inthat the ship is capable of breaking in the con- determining the coefficients.tinuous mode (i.e., at a speed of 3 knots [1.5 Milano has applied his method to numer-m/sl) is then given by ous ships of various shapes and sizes (e.g.,

Milano 1982) with reasonable correlation be-

hi= k 2- h/BA tween predicted and measured ice resistance.Milano also claims that his analytical modelis applicable to any hull form. Carter's (1982)

where B is the ship beam in meters, A is the comparisons between measured and predict-ship displacement in metric tonnes, and k2 is ed resistance are also quite satisfactory, inan empirical coefficient given by Tsoy as spite of his equation totally neglecting iner-

tial and buoyancy forces due to submergencek2= 0.031 to 0.035 of ice floes by the moving vessel. Further-

more, his treatment of the effect of ship veloci-

for icebreakers, and ty on ice resistance remains unconvincing tothis writer. Finally, his equation is based in

k 2 = 0.039- 0.025 part on elliptical approximations of the shipwaterlines, and therefore may not be readily

for ice-reinforced ships, with p being the ship applicable to the new, unconventional ice-block coefficient. breaker hull forms currently under develop-

ment.

18

hImaxl0n)

0 6 12 18 2414 I I I 7

- -- Tsoy (1983)

12 - Full-scale KTIKO 6Model

10 5

(knots) - (m/s)

6- 3

4 % -2

2 \ . 1.\-I

0 0.2 0.4 0.6hlrna x (Mn)

Figure 7 USCGC Katmai Bay performance curve inlevel ice measured during trials, predicted from modeltests, calculated by Tsoy's (1984) method.

CURRENT RESEARCH EFFORTS International cooperative researchIN ICE MODELING One of the tasks of the ITTC Ice Committee

is to develop standard methods for measure-Research and development efforts at ice ment of ice properties both in the field and in

modeling facilities are pursued in many di- the laboratory. It is also attempting to developrections: improvement of model ice, field guidelines showing which quantities need toand laboratory measurements or characteri- be documented during model tests and full-zations of ice properties, scale effects and test scale trials. The purpose of these standardsdata corrections, ice-propeller interaction, and guidelines is to allow meaningful corn-etc. International cooperative efforts between parison between results obtained at differentresearch facilities are also anderway under facilities. Obviously, they need to be periodi-the aegis of the ITTC Committee on Perfor- cally revised as progress in the understand-mance of Ships in Ice-Covered Waters (Ice ing of ship-ice interaction, testing methodol-Committee for short). ogy, and testing equipment is made.

At the outset it can be pointed out that be- The main recent effort of the Ice Committeecause of particular characteristics and condi- was the testing of a standard model by alltions (size, refrigeration system type and ca- participating ice tanks. The standard modelpacity, local environment, etc.) each ice tank selected was that of the Canadian Coasthas developed its own method for manufac- Guard "R"-class icebreaker. Fiberglas mod-turing a level ice sheet of prescribed mechan- els at the scales of 1:20 and 1:40 were con-ical properties. Modifications to seeding, structed by the National Research Council ofgrowth monitoring and tempering tech- Canada and circulated to the participants.niques (in the case of doped ice) are constant- The ice research facilities who participatedly introduced to improve the model ice consis- in the joint research program were IMD/NRCtency and repeatability. (Canada), CRREL (USA), HSVA (Germany),

19

WARC (Finland), AARI (USSR), JSRI and by direct Froude scaling of model tests resultsNEK (Japan). The test program (ITTC 1981) were consistently greater than those meas-included resistance and propulsion tests at ured during full-scale trials. From visualtwo ice thicknesses and two ice strengths. observation via underwater video and con-The results of the first series of tests were re- versation with icebreaker operators, in par-ported to the International Towing Tank ticular personnel of the U.S. Coast Guard, theConference in Goteborg, Sweden (ITTC 1984). conclusion was reached that ice ingestion byThis international cooperative research pro- the propellers is far greater in the model testsgram is continuing, the results of which are than at full scale. This conclusion is into be reported periodically to the ITTC. agreement with observations by Keinonen

(1983) who attributed this excessive ice-Ice testing propeller interaction to model ice-floes being

Testing for the mechanical properties of relatively larger than full-scale floes.ice is a difficult and time-consuming task. This author investigated the size distribu-Furthermore, all current methods give only tion of ice floes created by a simple wedgelocal values with the assumption that these (idealized icebreaker bow) towed through icevalues are representative of the whole ice sheets of varying thickness, flexuralsheet. Baker et al. (1987) have proposed an strength and strain modulus (Tatinclauxice-indexer for laboratory use that would 1986). It was found that the plan area andmeasure ice flexural strength, shear strength largest dimension of the floes followed a log-and crushing strength simultaneously and normal probability density function (Fig. 8).continuously. This ice-indexer is still under Similar results were presented by McKindradevelopment and evaluation, and Lutton (1981) from trial data with a USCG

While measurement of the friction coeffi- Great Lakes icebreaker (WTGB class). Thecient of ice is standard procedure in the labor- ratio of model floes' mean plan area to iceatory, no satisfactory field apparatus had thickness squared, A'hi, was found to be onlybeen devised. Kitagawa and Izumiyama proportional to the dimensionless strength,(1986) have recently reported on a portable af/yh, but independent of other parametersfield device for ice friction measurements such as the strain modulus or characteristicthat could be a significant addition to the tools length (Fig. 9). The coefficient of proportion-needed to characterize field ice. ality depended on the type of model ice (urea-

It is only recently that the fracture tough- doped or synthetic ice). The available full-ness of ice has received increasing attention scale data were insufficient to confirm theseas a material property significantly affect- last results.ing ship resistance in ice. As far as this writ- A subsequent study investigated the effecter knows, the Institute for Marine Dynamics of floe size on propeller torque (Tatinclauxof the National Research Council of Canada, 1987). Captive model propulsion tests wereSt. John's, Newfoundland, is the only ice made with the 1:10 model of the USCGC Kat-modeling facility where the fracture tough- mai Bay previously tested at CRREL. Theness of its model ice or some related index is tests were made in pre-sawn channels withmeasured routinely. No publication is yet floe sizes ranging from 1/2 to 1/6 the modelavailable on the relation between fracture beam, and in brash-ice-filled channels. At atoughness and ship model resistance, or on given advance coefficient, the torque coeffi-how to modify the full-scale prediction to in- cient Kq, increased linearly with floe sizeclude Kic. CRREL has recently initiated an (Fig. 10), and the torque standard deviationinvestigation of the fracture toughness of Sq increased with velocity, the rate of in-urea-ice (Bentley et al., in press), a study to be crease being some function of floe size. Also,expanded to freshwater ice. Standard meas- it appeared that below a certain ship speed nourement procedures and apparatus are yet to ice-propeller interaction occurred, and thatbe devised. this critical velocity decreased with increas-

ing floe size (Fig. 11).CRREL research on ship-ice interaction On the basis of the above study, it is suspect-

As was mentioned earlier, in this author's ed that model ice specific gravity and the ice-experience the power requirements predicted hull friction coefficient may significantly

20

SIMPLE WEDGE TESTS - TESTS 4 TO 6 COMIBINED SIMPLE WEDGE TESTS - TESTS 4 TO 6 COMBINED

20..0- 2 .8

10.0 8.4

• -. -. -,. N; N, PT ,,. t U "; " S S

L f A hI2

30.08

(Tatnclax 196).

20.0

8.i

10. St.D.- 1.4

0.08.

8.2

9.8w

N~~~~~MA PT 6.7 C T * Il

28.

4CTA"Kigoriak"in Syn Ice* "Kigoria'in Urea Ice0 Wedge in Urea Ice

2012

0 6"Kigoriok" in Syn Ice '

* "Klgoriak" (n Urea Icee e

A

L° 0b~h 0

00

00

i 0 100 200 300 400 5000 00 200 300 400 500 a-

-hi ), h ,

a. Floe area. b. Floe length.

Figure 9. Dimensionless test results (Tatinclaux 1986).

31

V= 50 cm/s

0.04 i -

_ ------- 0' I/

1SQ I 0

K (Nm)

0.02- BIce'

0

vs50Ocm/s In- 13rps I

0 20 40 0 20 40

Ice Floe Width (cm) Ice Floe Width (cm)

Figure 10. Effect of floe size on propeller performance (Tatinclaux 1987).

22

that ice will be ingested by the propeller at aBra c Foes ship speed lower than at full scale, and that it

E Aox4 cm o will also be more severe. An analytical and_ o Open Water $/ experimental study has been initiated to ad-

dress these questions. The results of a first0crude mathematical model of ice floe trajec-

2 / tory along a rigid surface are shown in Fig-o / ure 12.

, .. As a final comment, it is again empha-- • - C sized that one main difficulty in comparing

U) Jresults of model studies with full-scale re-- 2ptS suits is the lack of reliable and complete sets

- 8 of prototype data.

o 0 0NOVEL ICEBREAKING BOW DESIGNS

The past decade has witnessed the develop-0 20 40 60 ment of new icebreaking bows, the shapes of

V (CM /S) which depart radically from the conventionalones. A bow shape, first proposed and patented

Figure 11. Effect of floe size and velocity by Waas (1976), has been further developedon propeller torque standard deviation and extensively tested in West Germany both(Tatinclaux 1987). in the laboratory and in the field (Freitas

1981, 1982; Hellmann 1982; Freitas and Nish-izaki 1985; Schwarz 1986). From its original

iffect the trajectory of the ice floes along the conception (Fig. 13) the Waas bow hasiull and therefore the ice-propeller interac- evolved into radically new shapes (Fig. 14)..ion. As a ship (model or prototype) travels Basically, it has square shoulders, a very low.hrough ice, the broken floes are submerged stem angle at the water line, and an initially)y the bow to some depth, and they then rise be- flat bottom that evolves into a wedge. This!ause of buoyancy. Ice may be ingested if the bow first shears the level ice at the shoulders,;hip has traveled one ship-length before the then breaks it in bending into pairs of regu-loes have fully risen to near the surface. larly shaped ice floes, which are deflected3ince model ice is usually somewhat denser outwards under the remaining ice sheet. Ac-;han field ice, the model floes may be sub- cording to its developers, the main advantag-nerged deeper and rise slower than full- es of the Waas bow over conventional ice-;cale ice floes. Therefore, it can be expected breaker bows are: more efficient icebreak-

4 _ 1 1 _ _ I I I I __ -

- ° Y, / 2.

2 11- !2'z

3

Z 2 .

HUL COS-SITO" s. -0.90 0. 931.

0 10 20 30 4

Distance from Bow (Wn

Figure 12. Expected effect of ice density on ice floe trajectory alonga ship hull.

23

r -

Figure 13. Original Waas bow concept (Waas 1976).

Figure 14. Recent developments in Thyssen-Waas bow (FromTransactions of the ASME, Journal of Energy Resources,A. Freitas and R.S. Nishizaki, June 1986).

24

Figure 15. Wartsila experimental bow (From Enkvist and Mustamaki1986).

ing, i.e., reduced ice resistance and required adapt this bow to shallow draft icebreakers forpower; improved maneuverability in level operation in estuaries and rivers.ice; and creation of a nearly ice-free channel Another experimental icebreaking bowwith little or no ice-propeller interaction up to has been recently designed and tested at mod-speeds of 5 knots [2.5 m/s]. The effects of the el and full scales in Finland (Enkvist andWaas bow on a vessel's seakeeping capabili- Mustamaki 1986). This bow has a small stemties, performance in waves and ice ridges re- angle, circular waterlines and a plow or de-main a subject of controversy. A soviet ice- flector vanes to deflect the broken ice floesbreaker has recently been retrofitted with outwards (Fig. 15). Model and full-scale testssuch a bow, the U.S. Coast Guard has initiat- showed a significant decrease in icebreak-ed an evaluation program of its performance ing resistance in level ice over a convention-potential, and its developers are attempting to al bow, but no essential improvement of per-

25

formance in ridges and old broken channels, trials. Areas where improvements are moreand no improvement in maneuverability in urgently needed are:ice. An increase in open water resistanceand in slamming during transit in head sea 1. Knowledge of the effect of fracture tough.was observed. Enkvist and Mustamaki did ness on icebreaking resistance, breakingnot recommend this innovative bow for yes- pattern and ice floe size distribution.sels operating primarily in open water or inice ridges or rubble ice. They would limit the 2. Knowledge of the effect of ice floe sizeoperation of this bow to areas of thick level ice and model ice density on floe trajectory, ice-such as rivers, lakes and sheltered regions propeller interaction and propeller perfor-along a sea coast. In the oral presentation of mance (thrust, torque and shaft power).their paper, Enkvist and Mustamaki indicat-ed that the experimental bow was mounted on 3. Measurement of ice-hull friction coeffi-a towboat and used to open the Saima Canal to cient at model and full scales and its cffect onnavigation at the end of the 1985-86 winter. icebreaking resistance and propeller perfor-They reported that a navigation channel, mance.nearly free of ice, was opened in less than aday, when in previous years it had taken sev- 4. Analytical description of ship-ice inter-eral days for a towboat working alone to open action.the navigation channel, which contained alarge amount of brash ice. 5. Detailed documentation of full-scale

While not radically different from con- trials in ice, especially of the ice mechanicalventional icebreaker bows, a new bow was re- properties.cently designed by Melville Shipping Ltd, Ot-tawa, for the cargo ship MV Arctic. This 6. Development of apparatus and stan-Melville bow proved to significantly improve dardized methods for measuring ice proper-the ice performance of the MV Arctic (Baker ties, both at model and full scales; in particu-and Nishizaki 1986), confirming the model lar, an apparatus for field measurement oftests carried out at various facilities, ice-hull friction coefficients.

While this paper dealt only with modelingCONCLUSIONS AND FINAL REMARKS of ships in level ice, it should be mentioned

that research efforts are also actively pursuedSince the mid-1970s significant improve- in areas of ship motion in brash and mush

ments have been made in the methodology of ice, and of ship-ice-ridge interaction (ram-ship testing in ice for both EHP and SHP tests, ming tests) (ITTC 1984). The latter is of par-as well as in the development and characteri- ticular importance since extensive fields ofzation of model ice. The art and science of level ice are the exception in the arctic andmodel testing in ice has now reached the point subarctic regions, while ridges are everwhere model tank operators are confident present and can seldom be entirely avoided.that model test results can be used in predict-ing the performance of ice-going vessels, insuggesting modifications to existing ships to LITERATURE CITEDimprove their ice performance (e.g., Bakerand Nishizaki 1986), and in designing radi- Akkok, M., C.M. McCettles and S.J. Cala-cally new bow shapes for icebreakers and oth- brese (1986) Parameters affecting the kineticer ice-going vessels. The progress thus made friction of ice. The American Society of Me-has allowed the construction of ice-going yes- chanical Engineers, Transaction paper 86-sels with improved ice performance at lower Trib-16power requiremenits. Further improvements Ashton, G.D. (Ed.) (1986) River and Lake Iceare to be expected in the future as better under- Engineering. Littleton, Colorado: Water Re-standing ei" the physics of ship-ice interac- sources Publications.tion is gained and, especially, as more and Atkins, A.G. (1975) Icebreaking modeling.better field data are acquired from full-scale Journal of Ship Research, 19(1): 40-43.

26

-0-M"

Baker, D.N., G.W. Timco and C.C. Novicki Free, A.P. (in press) Solving icebreaker de-(1987) A strength indexer for model ice. In sign problems using ice model basins. Sub-Proceedings, 21st American Towing Tank mitted for publication to Marine TechnologyConference, Washington, D.C. National Society Journal.Academy Press, pp. 307-313. Freitas, A. (1981) Experimental investiga-Baker, D. and R. Nishizaki (1986) The MV tion of a novel icebreaker concept. In Pro-Arctic - New bow form and model testing. ceedings, Sixth Ship Technology and Re-Transactions, The Society of Naval Archi- search (STAR) Symposium, 17-19 June 1981,tects and Marine Engineers, 94: 57-74. Ottawa, Ontario. New York: The Society ofBentley, D.L, D.S. Sodhi and J.P. Dempsey Naval Architects and Marine Engineers, pp.(In press) Experimental determination of the 369-380.fracture toughness of urea model ice. In Pro. Freitas, A. (1982) New aspects of shipping inceedings, 10th International Conference on ice emergency from the expedition with thePort and Ocean Engineering under Arctic Max Waldeck. In Proceedings INTERMAR-Conditions (POAC '87), Fairbanks, Alaska. ITEC '84, 29-30 September 1982, Hamburg,Calabrese, S.J., R. Buxton and G. Marsh West Germany. Hamburg: Trimedia Wer-(1980) Frictional characteristics of materials beagentur, pp. 385-400 (in German with Eng-sliding against ice. Lubrication Engineer- lish. summary).ing, 36(5): 283-289. Freitas, A., R.S. Nishizaki (1985) Model testCarter, D. (1983) Ship resistance to continu- of an ice class bulk carrier with the Thyssen/ous motion in level ice. Transport Canada Waas bow form. In Proceedings, 4th Interna-Report No. TP-3679 E. Montreal, Quebec. tional Offshore Mechanics and Arctic Engi-Crago, WA., P.J. Dix and J.G. German neering (OMAE) Symposium, 17.21 February(1970) Model icebreaking experiments and 1985, Dallas, Texas. Vol. II. New York:their correlation with full scale data. Trans- American Society of Mechanical Engineers,actions, The Royal Institution of Naval Ar- pp. 330-335.chitects, London. Gow, A.J., H.T. Ueda and J.A. Ricard (1978)Edwards, R.Y., M. Dunne, G. Comfort, V. Flexural strength of ice on temperate lakes.Bulat and B. Johnson (1981) Results of full USA Cold Regions Research and Engineer-scale trials in ice with CCGS Pierre Radis- ing Laboratory, CRREL Report 78-9.son. In Proceedings, Sixth STAR Sympo- Hellmann, J.-H. (1982) Model and full scalesium, 17-19 June, Ottawa, Ontario. New tests in ice with the icebreaker Max Wal-York: The Society of Naval Architects and deck. In Proceedings, INTERMARITEC '82,Marine Engineers, pp. 291-310. 24-30 September 1986, Hamburg, West Ger-Enkvist, E. (1972) On the ice resistance en- many. Hamburg: Trimedia Werbeagentur,countered by ships operating in the continu- pp. 373-384.ous mode of ice breaking. Report No. 24. Hel- Hirayama, K (1983) Properties of urea-dopedsinki, Finland: The Swedish Academy of ice in the CRREL test basin. USA Cold Re-Engineering in Finland. gions Research and Engineering Laborato-Enkvist, E. and S. Makinen (1984) A fine- ry, CRREL Report 83-8.grain model ice. In Proceedings, IAHR Ice Hoffmann, L. (1985) Impact forces and fric-Symposium, 27-31 August, Hamburg. Vol. II, tion coefficient on the forebody of the Germanpp. 217-227. polar research vessel Polarstern. In Proceed.Enkvist, E. and E. Mustamaki (1986) Model ings, 8th International Conference on Portand full-scale tests with an innovative ice- and Ocean Engineering under Arctic Condi-breaker bow. Transactions, The Society of tions (POAC '85), 7-14 September 1985, Nars-Naval Architects and Marine Engineers, 94: sarssuag, Greenland. Vol. 3. Danish Hy-325-351. draulic Institute, pp. 1189-1202.Frederking, KL and G.W. Timco (1986) Field ITTC (1981) Report of the Committee on Per-measurements of the shear strength of colum- formance in Ice-Covered Waters. In Pro-nar-grained sea ice. In Proceedings, IAHR ceedings, 16th International Towing TankSymposium on Ice 1986, Iowa City, Iowa. Vol. Conference, 31 August - 9 September 1981, Le-1, pp. 279-292. ningrad, U.S.S.R. Vol. 1. Leningrad: Kry-

27

loy Shipbuilding Research Institute, pp. 363- July 1981, Quebec, Canada. Vol. I. Quebec:376. Universith Laval, pp. 235-243.ITTC (1984) Report of the Committee on Per- Mellor, M. (1983) Mechanical behavior of seaformance in Ice-Covered Waters. In Pro- ice. USA Cold Regions Research and Engi-ceedings, 17th International Towing Tank neering Laboratory, Monograph 83-1.Conference, 8-15 September 1984, Giteborg, Michel, B. (1979) Ice Mechanics. Les PressesSweden. Vol. 1. SSPA, G6teborg, pp. 584-625. de I'Universite Laval, Quebec.Jones, N. (1986) A note on ice scaling. Jour- Michailidis, M. and D.C. Murdey (1981) Per-nal of Ship Research, 30(2): 134-135. formance of CCGS Franklin in Lake Mel-Kashtelyan, V.I., I.I. Poznyak and A.Y. ville, 1980. In Proceedings, Sixth STAR Sym-Ryvlin (1968) Ice resistance to motion of a posium, 17-19 June, Ottawa, Ontario. Newship. Translation by Marine Computer Ap- York: The Society of Naval Architects andplication Corporation (1969), from Soprotivle- Marine Engineers, pp. 311-322.nie L'da Dvizheniyu Sudna, Sudostroenie, Milano, V.R. (1973) Ship resistance to contin-Leningrad. uous motion in ice. Transactions, Society ofKayo, Y., et al. (1983) Field study on me- Naval Architects and Marine Engineers, 81:chanical strength of sea ice at east coast of 274-299.Hokkaido. In Proceedings, 7th International Milano, V.R. (1982) Correlation of analyticalConference on Port and Ocean Engineering prediction of ship resistance in ice with mod-under Arctic Conditions (POAC '83) , 5-9 el and full scale test results. In Proceedings,April, Helsinki, Finland. Vol. I, pp. 109-118. INTERMARITEC '82, 24-30 September, Ham-Keinonen, A. (1983) Major scaling problems burg, West Germany, pp. 350-372.with ice model testing of ships. In Proceed- Oksanen, P. (1983) Friction and adhesion ofings, Twentieth American Towing Tank ice. Technical Research Center of Finland,Conference, 2-4 August, Hoboken. New Jer- Publication 10.sey. Vol. II, pp. 595-612. Parsons, B.L., J.B. Snellen and B. HillKitagawa, H. and K. Izumiyama (1986) An (1986) Physical modeling and the fractureexperience with a portable device for ice fric- toughness of sea ice. In Proceedings, Fifthtion measurements. Communication to ITTC International Offshore Mechanics and Arc-Committee on Ships in Ice-Covered Waters tic Engineering (OMAE) Symposium, 13-18(unpublished). April, Tokyo, Japan. Vol. IV. New York:Kotras, T.V., A.V. Baird and J.N. Naegle American Society of Mechanical Engineers,(1983) Predicting ship performance in level pp. 358-364.ice. Transactions, The Society of Naval Ar- Ryvlin, A.Y. (1973) Experimental study ofchitects and Marine Engineers, 91: 329-349. the friction of ice. In Ice Navigation Quali-Lavrov, V.V. (1969) Deformation and ties of Ships (D. Kheisin and Y.N. Popov,strength of ice. National Science Foundation, Eds.) Translation published as USA Cold Re-Translation TT 70 50120. gions Research and Engineering LaboratoryLewis, J.W., F.W. DeBord and VAL Bulat Draft Translation 417, pp. 217-234.(1982) Resistance and propulsion of ice- Saeki, H., 0. Toshiyuki, N. Nakazawa andworthy ships. Transactions, The Society of T. Sachito (1984) The coefficients of frictionNaval Architects and Marine Engineers, between sea ice and various materials usedNew York, 90: 249-276. in offshore structures. In Proceedings, 16thLiukkonen, S. (1985) Friction measurements Annual Offshore Technology Conference, 7-9in full scale and in model scale and the mod- May, Houston, Texas. Vol. 1, pp. 375-382.el surface treatment method used at WARC. Schultz, L.A. and A.P. Free (1984) Recent ex-Report to the ITTC Committee on Ships in Ice- perience in conducting ice model tests usingcovered Waters (unpublished). a synthetic ice modeling material. In Pro-McKindra, C.D. and T.C. Lutton (1981) Sta- ceedings, IAHR Ice Symposium, 27-31 Au-tistical analysis of broken ice dimensions gust, Hamburg, West Germany. Vol. II, pp.generated during 140-ft WTGB icebreaker 229-239.trials. In Proceedings, 6th International Schwarz, J. (1975) On the flexural strengthConference on Port and Ocean Engineering and elasticity of saline ice. In Proceedings,under Arctic Conditions (POAC '81), 27-31 Third IAHR International Symposium on Ice

28

Problems, 18-21 August, Hanover, New Engineering Laboratory, CRREL Report 85-Hampshire, pp. 373-386. 4.Schwarz, J. (1977) New developments in mod- Tatinclaux, JC. (1986) Ice floe distributioneling ice problems. In Proceedings, 4th Inter- in the wake of a simple wedge. In Proceed-national Conference on Port and Ocean En- ings, Fifth International Offshore Mechan-gineering under Arctic Conditions (POAC ics and Arctic Engineering (OMAE) Sympo-77), St. John's, Newfoundland, Vol. I, pp. 45- sium, 13-18 April, Tokyo, Japan. Vol. IV, pp.61. 622-629.Schwarz, J. (Ed.) (1979) IAHR Recommenda- Tatinclaux, J.C. (1987) Effect of ice-floe sizetions on testing methods of ice properties. on propeller torque in ship-model tests. InUniversity of Lulea, Sweden, Research Re- Proceedings, 21st American Towing Tankport Tulea 1979:03. Conference, Washington, D.C. NationalSchwarz, J. (1983) Modeling techniques in Academy Press, pp. 291-298.various ice model basins. WEGEMT 1983, Tatinclaux, J.C., KA. Forland and D. Mur-Lecture XIV. dey (1986) Laboratory and field studies of iceSchwarz, J. (1986) Advances in icebreaker friction coefficient. In Proceedings, IAHRtechnology in West Germany. In Proceed- Symposium on Ice, 18-22 August, Iowa City,ings, International Polar Transportation Iowa, Vol. I, pp. 389-400.Conference (IPTC '86), 4-8 May 1986, Van- Timco, G.W. (1979) The mechanical andcouver, B.C. Vol. One. D.F. Dickins Asso- morphological properties of doped ice: Aciates Ltd., pp. 201-219. search for a better structurally simulated iceSchwarz, J. and W.F. Weeks (1977) Engi- for model test basins. In Proceedings, 5th In-neering properties of sea ice. Journal of Gla- ternational Conference on Port and Oceanciology, 19(81): 499-531. Engineering under Arctic Conditions (POACSodhi, DS., K Kato, F.D. Haynes and K. Hi- 79) 13-18 August, Trondheim, Norway, Vol.rayama (1982) Determining the characteris- 1, pp. 719-739.tic length of model ice sheets. Cold Regions Timco, G.W. (1980) The mechanical proper-Science and Technology, 6: 99-104. ties of saline-doped and carbamide (urea)-Sodhi, D.S., C.R. Martinson and W.B. doped model ice. Cold Regions Science andTucker LU (1985) Determining the character- Technology, 3: 45-56.istic length of floating ice sheets by moving Timco, G.W. (1981a) A comparison of sever-loads. In Proceedings, Offshore Mechanics al chemically doped types of model ice. Inand Arctic Engineering, 17-21 February, Proceedings, IAHR Ice Symposium, 27-31Dallas, Texas. Vol. II. New York: American July, Quebec, Vol. II, pp. 489-502.Society of Mechanical Engineers, pp. 155- Timco, G.W. (1981b) On the test methods for159. model ice. Cold Regions Science and Tech-Tabata, T., K. Frijino and M. Aota (1967) nology, 4: 269-274.Studies of the mechanical properties of sea Timco, G.W. (1985) Flexural strength andice. XI. The flexural strength of sea ice in fracture toughness of urea model ice. ASMEsitu. In Physics of Snow and Ice, Vol. 1. Part Journal of Energy Resources Technology,I. Sapporo, Japan: Institute of Low Tempera- 107:498-505.ture Science, Hokkaido University, pp. 539- Timco, G.W. (1986) EG/AD/S: A new type of550. model ice for refrigerated towing tanks. ColdTatinclaux, J.C. (1984a) Model tests on two Regions Science and Technology, 12: 175-models of WTGB 140-foot icebreaker. USA 195.Cold Regions Research and Engineering La- Timco, G. W. and R.M.W. Frederkingboratory, CRREL Report 84-3. (1982) Comparative strengths of freshwaterTatinclaux, J.C. (1984b) Model tests in ice of ice. Cold Regions Science and Technology,a Canadian Coast Guard R-class icebreaker. 6: 21-27.USA Cold Regions Regions Research and Timco, G.W. and R.M.W. FrederkingEngineering Laboratory, Special Report 84-6. (1983) Flexural strength and fracture tough-Tatinclaux, J.C. (1985) Propulsion tests in ness of sea ice. Cold Regions Science andlevel ice on a model of a 140-ft WTGB ice- Technology, 8: 243-253.breaker. USA Cold Regions Research and

29

Timco, G.W. and R.M.W. Frederking ical Engineers, pp. 178-182.(1986) Confined compression tests: Outlining Vance, G.P, AS. Gracewski and M.J. Good-

the failure envelope of columnar sea ice. Cold win (1981) Full scale icebreaking tests of the

Regions Science and Technology, 12: 13-28. USCGC Katmai Bay. In Proceedings, Sixth

Tryde, P. (Ed.) (1979) Physics and mechan- STAR Symposium, 17-19 June, Ottawa, Onta-

ics of ice. In Proceedings of IUTAM Sympo- rio. New York: The Society of Naval Archi-

sium, 6-10 August, Copenhagen. Berlin: tects and Marine Engineers, pp. 323-343.Springer-Verlag. Vaudrey, K.D. (1977) Determination of me-Tsoy, L.G. (1983) Modeling of vessel's move- chanical sea ice properties by large-scalement in a channel broken up by icebreaker. field beam experiments. In Proceedings, 4thIn Proceedings, 7th International Conference International Conference on Port and Oceanon Port and Ocean Engineering under Arctic Engineering under Arctic Conditions (POACConditions (POAC '83), 5-9 April, Helsinki, '77), St. John's, Newfoundland. Vol. 1, pp.Finland, Vol. II, pp. 654-663. 529-543.Urabe, N. and A. Yoshitake (1984) Similitude Waas, EL (1976) Ice-breaker vessel. U.S. Pat-

law of ice based on fracture mechanics. In ent No. 3,984,091.Proceedings, Third International Offshore Weeks, W.F. and A. Assur (1967) The me-

Mechanics and Arctic Engineering (OMAE) chanical properties of sea ice. USA Cold Re-

Symposium, New Orleans, Louisiana. Vol. gions Research and Engineering Laborato-I1. New York: American Society of Mechan- ry, Monograph II-C3.

30 U S GOVERNMENT PRINTING OFFICE 1988-600-057/82040

A facsimile catalog card in Library of Congress MARC format is repro-duced below.

Tatinclaux, Jean-Claude P.Ship model testing in level ice: An overview / by Jean-Claude P. Tatin-

claux. Hanover, N.H.: U.S. Army Cold Regions Research and EngineeringLaboratory; Springfield, Va.: available from National Technical Informa-tion Service, 1988.

v, 39 p., illus.; 28 cm. (CRREL Report 88-15.)Bibliography: p. 26.1. Model ice. 2. Model tests. 3. Ship models. 4. Towing tank tests.

1. United States Army. Corps of Engineers. II. Cold Regions Research andEngineering Laboratory. Ili. Series: CRREL Report 88-15.