H. S.

e-mail: hs-john.kim@samsung.com

Speed-Power Performance of 95,000DWT Arctic Tanker Design
When Arctic offshore development in the 1970s first led to consideration of ice capable tankers, there was a high level of uncertainty over design requirements for both safety and ship performance, and a lack of reliable methods to evaluate design proposals. Since that time, improved understanding of the ice environment has raised the confidence of design specifications. Parallel developments have resulted in a suite of engineering tools for ship performance evaluation at the design stage. Recent development of offshore and near shore oil and gas reserves in several countries, together with economic studies of increased transportation through the Russian Arctic, led to renewed interest in ice capable tanker design. In response, Samsung Heavy Industries (SHI) applied its experience in tanker design and construction to the design of a specialized tanker with ice capability. SHI produced two prototype hull designs for further study. The performance of both hulls and of the propellers was evaluated at the Institute for Ocean Technology (IOT) in St. John’s, Newfoundland. This paper discusses the development of the design, describes the model experiments to determine performance and variations, and presents the results. It shows how physical modeling can provide insight into design features, and points out the areas where further research will have the greatest effect. DOI: 10.1115/1.1894406

M. K. Ha
Samsung Heavy Industries, Koje-City, Korea

F. M. Williams D. Molyneux
e-mail: David.Molyneux@nrc-cnrc.gc.ca NRC Institute for Ocean Technology, St. John’s, Newfoundland, Canada

H. H. Chun
e-mail: chunahh@pusan.ac.kr Pusan National University, Pusan, Korea



SHI undertook a study leading to the design of a tanker operating on a hypothetical route between Western Europe and the Dikson area. SHI designed and built the Hibernia oil platform shuttle tankers, which operate in occasional pack ice. Design requirements for the proposed Dikson route, where there would be continuous operation in heavy ice, were much more severe. Ice along the voyage route is mainly first year level ice, with thickness less than 2 m, and midwinter strength about 500 kPa. A typical operating scenario for a target region of interest would require the tanker to maintain a reasonable speed in 1 m of medium first year ice, and to be assured of making progress in 2 m of first year ice. In practice, the tanker must also be able to operate in snow-covered ice, and be capable of transiting certain ridge and rubble configurations. Furthermore, the tanker must have a verifiable turning ability in the specified ice conditions. These requirements were considered in the overall performance evaluation, but this paper concentrates on the level ice power requirements.

This, in turn, placed restrictions on the brake horsepower that could be delivered. SHI selected a twin skeg design to meet the requirements of the general arrangement and also to provide protection for propeller shafts in heavy ice. The angle, inclination, and separation distance for the skegs were developed design parameters. 2.2 Speed-Power Prediction. At the concept stage, a variety of methods may be used to estimate ship power requirements. Normally the estimate combines a resistance calculation with propulsive efficiency. The ice resistance Ri depends on ship velocity v, length L, draft T, and beam B, as well as ice thickness h, ice strength f , water and ice densities w and i, respectively, and the buoyant density = w − i. Studies of full scale and model scale data 1–3 resulted in the equation R/
wgBh 2

= C0 + C1 f /


+ C2v/ gh



Icebreaking Tanker Concept Design

2.1 General Requirements. SHI selected a 95 k DWT Aframax class tanker as the basis for the design. The displacement is 119 k tons, with beam, length and block coefficient constraints of 44 m, 270 m, and 0.749, respectively. Limited water depth on portions of the route restricted tanker draft to 11.5 m. The ship is required to break 2 m of 500 kPa level ice, and maintain speeds of 16 knots in open water and 4 knots in ice. The objective of the concept design was good performance in both open water and ice. The design study began with a review of information on the performance of different icebreaking bow types. SHI selected the Canadian R-Class bow, a classical icebreaking bow with relatively low wave resistance and good sea-keeping characteristics, as the basis for the bow design. There is little information in the literature on stern design, and yet it is a critical factor. The 11.5 m draft restriction and necessary propeller tip clearance limited propeller diameter less than 7 m.
Contributed by the OMAE Division for publication in the JOURNAL OF OFFSHORE MECHANICS AND ARCTIC ENGINEERING. Manuscript received February 26, 2004; final revision, October 9, 2004. Review conducted by: S. Jones.

where the coefficients C0, C1, and C2 are given by empirical formulas. Through dimensional analysis, Vance 4 derived an equation similar to 1 with the addition of buoyant density and ship length. Theoretical determinations of resistance that are based on detailed calculations for the ice breaking forces, such as Naegle 5 and Milano 6 , take into account hull form factors but idealize the ship-ice interaction. Lindqvist’s method 7 gives consideration to ice properties and is reasonably simple to apply. Figure 1 shows sample calculations of ice resistance at ship speed of 1.544 m / s 3 knots using the methods of Edwards et al. 1 and Lindqvist 7 . The propulsive efficiency of a ship at icebreaking speeds is generally much lower than the propulsive efficiency at its design speed in open water. The very nature of screw propeller operation means that the highest propulsive efficiencies typically between 0.65 and 0.7 are obtained at advance ratios much higher than are practical for icebreaking speeds. As a result propulsive efficiency at icebreaking speeds can be as low as 0.11. The propulsive efficiency can be further reduced by the amount of ice flowing through the propeller, which results in obstructed flow and mechanical impact with the blades 8 . The spread in resistance estimates, combined with the difficulty in predicting the effect of ice on propulsive efficiency makes theoretical power predictions unMAY 2005, Vol. 127 / 135

Journal of Offshore Mechanics and Arctic Engineering Copyright © 2005 by ASME

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 1 Estimates of ice resistance, 3 knots.

satisfactory even at the concept design stage. Hence SHI relied on the model tests of the initial designs for power estimates.


Hull Form Development

3.1 Initial Lines of Development. The lines shown in Fig. 2 a was the first attempt at meeting the performance specifications within the constraints of draft and displacement discussed above. Stem angle A, ice knife B, shoulder angle C, and flare angle D were key design parameters. The ice knife aids in devel-

opment of the stem crack and prevents the ship from riding up too far on the ice 9 . Small stem and shoulder angles reduce the horizontal component of the icebreaking force 10 , while increasing angles promote clearing of ice away from propellers. For the tanker, the stem angle is 25 deg at the waterline and 35 deg near the knife, and the shoulder angle is 30 deg. To meet the highest ice classification requirements within the Russian Class Rules, a flare angle of 8 deg was included in the mid-body to reduce the likelihood of entrapment in heavy ice 11 . At the stern, inward, outward, and vertical skegs were considered. The commercial code SHIPFLOW was used to analyze the flow field. Figure 3 shows the hull form grid generation and streamlines on a twin skeg stern. Jonk 12 studied the effect of vertical inclination of the skeg on the propulsive performance of the ship. He concluded that the difference in propulsive performance was caused by the pressure difference between the inside and outside of the skeg and the cross flow induced by this difference. The results in Fig. 3 show that the outward inclination will induce ice pieces to flow to the outside. A skeg with 3 deg outward angles “K” of Fig. 2 , which also has the benefit of increased open water flow into the propeller, was adopted. The distance between the skegs of SHI’s tanker is 0.39B and the inclination of the skeg bottom is 13 deg to 15 deg. The propeller

Fig. 2

„a… M493, Body Plan and Bow Profile; „b… M501, Body plan and bow profile

Fig. 3 The flow pattern of three skeg direction types; „a… inward skeg, „b… straight skeg, „c… outward skeg

136 / Vol. 127, MAY 2005

Transactions of the ASME

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Table 1 Principal dimensions of M501 Items LBP m Breadth m Depth m Draft m Cb Displacement cu. m Values 270 44 20 11.5 0.7499 102,458

rotation direction is outward. Outward rotation deflects oncoming ice flow away from the propeller but is unfavorable to the propulsion performance in the open sea. 3.2 Hull Lines Modification. The first set of model tests, in ice and open water, were carried out at the ice basin of IOT, NRC-CNRC, St. John’s Newfoundland, Canada. Hull M493 had acceptable open water performance but the ice clearing resistance component was higher than ice breaking resistance and the propulsive efficiency in ice was considerably lower than the equivalent open water value. In level ice, large numbers of ice pieces reached the propeller disk area. Ice pieces impacting the propeller blades increased torque and reduced propulsive efficiency. The main intention of the hull modification was to reduce the inflow of ice pieces to the propeller disk area and to improve the ice clearing performance. The enhanced flow generated by the outward skeg also increased the amount of ice directed towards the propeller. Hence the skeg angle “K” of Fig. 2 was changed to inward to reduce the flow between the skegs. To reduce the size of the ice pieces, and also to reduce the distance ice pieces moved along the hull, the bow angles “C” of Fig. 2 were increased. It was anticipated that this change would reduce ice clearing and buoyancy forces but increase breaking force because of the additional energy absorbed in breaking smaller pieces. Third, the buttock angle “M” of Fig. 2 was reduced to improve ice clearing. Figure 2 b shows the modified hull form M501. Table 1 gives the principal dimensions.

Fig. 4 Comparison of effective power in open water, full scale

4.2 Ship Open Water Performance. For the open water tests with M493, the water temperature was 0.5 ° C and the kinematics viscosity was 1.76 10−6 m2 / s. With M501, the water temperature was 3.2 ° C and the kinematics viscosity was 1.61 10−6 m2 / s. Model values were converted to full scale for salt water at 15 deg using form factors 1 + K of 1.23 and 1.30, respectively. Figure 4 compares the effective powers, calculated from resistance, over the open water speed range. The higher values for M501 are due to slight increases in wetted surface and displacement. 4.3 Resistance in Ice. Provided that the wave making component of the hydrodynamic resistance is small, resistance in ice can be assumed to be the sum of hydrodynamic, ice breaking, ice submergence, and ice clearing components 15 ; Rtot = Row + Rs + Rcl + Rbr 2 The submergence, clearing, and breaking forces for any speed and ice thickness may be expressed in dimensionless form by Rs = Cs gBhd gh v2
fh iB v 2 c/2

3 4

Rcl = Ccl iBhv2 Rbr = Cbr iBhv2


Physical Model Experiments


4.1 Description of Experiments. The ship models were constructed of High Density foam coated with fiberglass and finished with epoxy paint. A scale of 1:31.938 was selected so that stock propellers could be used for the initial tests. The models were ballasted to specified draft, GM, and radii of gyration. The EGADS CD model ice in the ice tank at IMD provides the kinematics and mechanical characteristics required to model the ship-ice interaction correctly 13,14 . Model ice thicknesses were selected to represent 1 m and 2 m thick ice at full scale. For each ice sheet, ice thickness, density, flexural and shear strengths, compressive failure stress, and ship-ice friction coefficient were measured. For resistance tests, the fully appended model was fitted with dummy propeller hubs. The model was fixed to the tow carriage, free in pitch, heave, and roll. For self-propulsion tests, M493 was fitted with IMD stock propellers 66L and 66R, while M501 was fitted with Samsung designed propellers. Thrust, torque and shaft rotation speed were measured separately on each shaft. For the self-propulsion experiments, the model was towed through the ice at constant speed, and propeller revolution was varied to give positive and negative values of the tow force. The model selfpropulsion point was interpolated from the data as the point at which the tow force was zero. Maneuvering experiments were carried out with both hulls, and a wake survey was completed for Model 501, which was fitted with the Samsung propeller. This paper concentrates on the resistance and self-propulsion results. Journal of Offshore Mechanics and Arctic Engineering

The component force equations are similar to those derived in 15 , verified with a large data set for the Canadian R-Class icebreaker. In the current study, the nondimensional coefficients c, b, Cs, Ccl, and Cbr were determined for each hull form from the model tests according to the procedure outlined below. Model resistance was measured in open water, and in level ice and presawn ice at two ice thicknesses. Hydrodynamic resistance Row was determined as a second order function of velocity from the open water results. The resistance in presawn ice was identified as Row + Rs + Rcl, from which Row was removed. When Rs + Rcl was plotted against v at constant h, the intercept was Rs. The dimensionless Rcl was then plotted against the thickness Froude number, gh / v2, for all v and h, to determine Ccl and c. Calculated values of Row + Rs + Rcl, were subtracted from the total resistance in level ice to give Rbr. Finally, the dimensionless Rbr was plotted against the strength number, f h / iBv2, for all v and h, to determine Cbr and b. The coefficients for the two hull forms are given in Table 2. From the nondimensional coefficients, the icebreaking resistance in specified conditions at any scale may be determined. Figure 5 shows the model resistance forces for M501in 63 mm level ice. The comparison of results showed that the largest difference was in the breaking component. The total breaking force was higher for M501, as anticipated in the design modification. Contrary to expectations, the submergence and clearing forces were not reduced for M501. Over the range of velocities considered, the MAY 2005, Vol. 127 / 137

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Table 2 Comparison of resistance coefficients Cs M493 M501 1.10 1.15 Ccl 1.86 2.18 c 1.02 0.96 Cbr 0.89 1.59 b 1.92 1.67

differences were small. Figure 6 compares breaking and clearing resistance components over a range of model speeds for M493 and M501 in 63 mm level ice. 4.4 Power Requirements in Ice. The measured parameters in the model self-propulsion tests in ice were expressed as quadratic functions of propeller speed at each model speed in each set of ice conditions. The self-propulsion point was determined as the propeller speed at which the model tow force vanished. Total thrust and total torque, expressed as quadratic functions of propeller speed, were then evaluated at the self-propulsion point. For analysis of the self-propulsion in ice, it is convenient to refer to coefficients, such as hull factors and propulsive efficiency, used in analysis of self-propulsion in open water. These parameters may be used to determine relative performance in ice and open water. However, they do not correctly represent the dynamics of a propeller in the presence of ice, since the changes in propeller force due to ice impacts or blocked flow may be much larger than the effects of hull and wake in normal operating conditions. For example, wake fraction uses thrust or torque identity between the propeller in open flow and the propeller behind the ship to determine an equivalent mean flow velocity for the propeller behind the ship. When ice is introduced the flow can no longer be steady and the concept of a wake fraction, based on the mean performance of the propeller behind the ship, no longer applies. Relative rotative efficiency is also a link between a mean steady flow in open water and behind the ship, and this concept is

also no longer valid in ice. Thrust measured in ice can be compared with the resistance in ice, and a thrust deduction coefficient determined. The equivalent coefficient can be determined from the open water overload experiments. For the open water values, we have chosen to use a flow identity method, which considers the open water overload data at the same model speed and shaft rate of rotation as the self-propulsion point in ice. In this case, the only hydrodynamic difference is the broken ice interfering with the propeller. Overall propulsive efficiency is another valid parameter, provided that the main factors affecting propulsive efficiency in ice are correctly modeled in the ice tank. These are the size, strength, density and velocity of ice pieces, and the flow of ice around the stern. Hence the procedure for ship performance prediction assumes that the net propulsive efficiency value obtained for the model applies to the ship. Based on the results of model experiments, the propulsive efficiency is primarily a function of ship speed and ice thickness. So, when calculating delivered power for ice strengths other than the ones used for testing, it is assumed that the propulsive efficiency does not change. For each model speed and ice condition, actual resistance was calculated from the results of the experiments. Then the thrust deduction coefficient t was determined from Eq. 6 and an effective net propeller efficiency D in ice was decided by Eq. 7 , t=1− Rtot T 6



Rtotv 2 nQ


where, T, Q, and n are thrust, torque, and shaft speed, respectively. For comparison, the same calculations were made with the open water results at the overload conditions, that is at the same model and propeller speeds: t=1− and

Rtot + Fx T



Rtot + Fx v 2 nQ


Fig. 5 Model scale resistance components for M501, 63 mm

Fig. 6 Model scale resistance components for M493 and M501, 63 mm ice

where, Fx is the tow force at the given shaft speed. The open water efficiency is the optimum for the given overload condition. Table 3 contains these results. Figure 7 shows a comparison of the overall propulsive efficiency for the two models in 63 mm and 30 mm of level ice, respectively. The change in flow pattern near the stern due to the change in skeg angle did not have a significant effect on propulsive efficiency. A small effect was observed at the highest speed. Thrust deduction is a coefficient that can be applied to propulsive performance in ice, but its interpretation requires some care. Ice impacts with the propeller are recorded as loss of thrust, but typically the ice impacts are very short duration. If the ice impacts are very large, they will influence the mean value of thrust. If the model ship was self-propelled, it would slow down as a result of the impacts. However, since the model is towed at constant speed, the average thrust value can appear to be less than the value needed to move the ship. Negative values of thrust deduction factor in ice are an indication of the degree of propeller ice interaction, rather than the thrust needed to propel the ship. The negaTransactions of the ASME

138 / Vol. 127, MAY 2005

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Table 3 Model propulsive efficiency in open water and in ice t V m/s h = 30 mm 0.18 0.27 0.36 0.55 h = 63 mm 0.18 0.27 0.36 0.55 M493 ice 0.20 0.31 0.22 −0.15 −0.05 −0.06 t M493 ow 0.07 0.09 0.10 0.06 0.08 0.09 d M493 ice 0.12 0.20 0.32 0.12 0.17 0.23 d M493 ow 0.16 0.23 0.31 0.11 0.18 0.25 −0.13 −0.09 −0.06 0.13 0.15 0.17 0.16 0.18 0.19 0.14 0.16 0.19 −0.16 0.04 0.04 0.25 0.26 0.33 0.23 0.21 0.30 0.24 0.24 0.27 t M501 ice t M501 ow d M501 ice d M501 ow

tive values of t ice for both hulls in 63 mm ice indicate frequent propeller ice impacts, an observation that was corroborated by video records of the experiments. In thin ice, M493 experienced few impacts, hence t ice is high. For M501, ice from the side of the hull would occasionally become trapped in the propeller stream, and so the value of t ice are mixed. Changes in thrust and torque can be used as a simple measure of the degree of propeller-ice interaction. The ratio of Ti/ To in Fig. 8 gives a measure of the effect of ice on the thrust and the ratio of Qi/ Qo in Fig. 8 gives a measure of the effect of ice on the torque. These ratios for M493 and M501 are compared in Fig. 8. These figures show that both models have similar values of Ti/ To, but M501 has a higher value of Qi/ Qo throughout the speed range. The values obtained from M493 and M501 compare well with other ships 8 . Ship resistance in level ice at full scale is the sum of the components in Eq. 1 . The open water resistance, Row, was determined from the model scale open water resistance by scaling the wave making and frictional contributions separately, in the tradi-

tional manner. Table 3 shows these calculations. Skin friction was determined from Reynolds Number using the ITTC 1957 line. A form factor of 1.3 and a correlation allowance of 0 were applied. The components of ice resistance may be calculated using the relations in Eqs. 2 – 5 , the coefficients in Table 2, and the fullscale values of ship and ice parameters. Delivered power was calculated from effective power using the propulsive efficiencies in ice reported in Table 3. Figure 9 shows a comparison between M493 and M501 in 2 m of level ice with a flexural strength of 500 kPa. This shows that M493 has a slightly lower effective power than M501, although in absolute terms this difference is small. When the propulsion is included, the difference becomes large at 6 knots.

Fig. 8 Propulsion ratios in ice, based on flow identity

Fig. 7 Propulsive efficiencies, 63 mm ice „top…, 30 mm ice „bottom…

Fig. 9 Comparison of power predictions in 2 m level ice

Journal of Offshore Mechanics and Arctic Engineering

MAY 2005, Vol. 127 / 139

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm



This work was conducted in cooperation with Samsung Heavy Industries Ltd. Korea , ASERC of Pusan National University Korea , MUN Canada , and NRC-CNRC IOT Canada .

Small differences in hull form parameters can be related to specific differences in vessel performance in ice. Increased bow angles resulted in an increase in the breaking component of resistance, with larger differences at higher speed and thicker ice. The reason for the slight increases in submergence and clearing components was determined by examining video records of the model experiments. Steeper waterline angles meant that some ice pieces were pushed to the side rather than under the bow, and there was occasional catching and jamming of those pieces at the shoulders. This increase in the clearing force, occurring at all speeds, may have shown up in the submergence component as well. The change in flow pattern near the stern due to the change in skeg angle did not have a significant effect on propulsive efficiencies. A small effect was observed at the highest speed, with the efficiency of M501 being lower. The propulsion ratios in Fig. 8 suggest more propeller-ice interaction for M501. Ice caught between the hull and the side of the channel, and then swept into the propeller, is the likely explanation. The dominant feature of the stern design is the twin skeg. Underwater video records show that ice pushed down under the bow may travel along the bottom and then be guided between the skegs towards the propeller disks. Careful analysis of the video does suggest hull form modifications, which might alleviate this effect, which, in turn, would improve propulsive efficiency. As Fig. 9 indicates, the greatest improvement in ship performance can be achieved by increasing propulsive efficiency in ice. This project pushed the conventional boundaries in tanker design and exemplified the trade-off between hull form constraints and high performance demands imposed by the operating conditions. Two hull forms were designed and tested. The physical model tests showed clearly the effect of hull form parameters on performance, and pointed the way to design improvements. Full scale performance data for tankers operating in the design ice conditions remains a vital requirement to complement design studies such as this one 16 .

1 Edwards Jr., R. Y., Major, R. A., Kim, J. K., German, J. G., Lewis, J. W., and Miller, D. R., 1976, “Influence of Major Characteristics of Icebreaker Hulls on Their Powering Requirements and Maneuverability in Ice,” Transactions of SNAME, 84, pp. 364–407. 2 Lewis, J. W., DeBord, F. W., and Bulat, V. A., 1982, “Resistance and propulsion of ice-worthy ships,” SNAME Transactions, 90, pp. 249–276. 3 Lewis, J. W., and Edwards, R. Y., 1970, “Methods for predicting icebreaking and ice resistance characteristics of icebreakers,” SNAME Transactions, 78, pp. 213–249. 4 Vance, G. P., 1975, “A scaling system for vessels modelled in ice,” IceTech 75, Montreal. 5 Naegle, J. N., 1980, “Ice resistance prediction and motion simulation for ships operating in the continuous mode of icebreaking,” Ph.D. thesis, University of Michigan, Ann Arbor. 6 Milano, V. R., 1980, “A reanalysis of ship resistance when in continuous motion through solid ice,” Intermaritec Symposium, Hamburg, pp. 456–475. 7 Lindqvist, G., 1989, “A straightforward method for calculation of ice resistance of ships,” The 10th International Conference on Port and Ocean Engineering under Arctic Conditions, Luleå, pp. 722–735. 8 Molyneux, D., Simoes Re, A., Spencer, D., and Reynolds, A., 1990, “Recent developments in model experiment techniques for large icebreakers,” IceTech90, Calgary. 9 White, R. M., 1970, “Prediction of icebreaker capability,” Transactions of RINA, 112 2 , pp. 225–251. 10 Schwarz, J., 1986, “Some latest developments in icebreaker technology,” J. Energy Resour. Technol., 108, pp. 161–167. 11 Russian Maritime Register of Shipping, 1995, “Rules for the classification and construction of sea-going ships,” Russian Maritime Register of Shipping. 12 Jonk, A., 1985, “The use of Non-Viscous Flow Calculations in Hull Form Optimization,” Workshop on Development in Hull Form Design 1, MARIN. 13 Spencer, D. S., and Timco, G. W., 1990, “CD Model Ice: a process to produce correct density CD model ice,” IAHR Tenth International Symposium on Ice, Espoo, Finland, 745–755. 14 Timco, G. W., 1986, “EG/AD/S: A new type of model ice for refrigerated towing tanks,” Cold Regions Sci. Technol., 12, pp. 175–195. 15 Spencer, D., Williams, F. M., and Newbury, S., 1992, “Model scale/full scale ice resistance correlation for the CCG R-Class icebreakers,” LM–1992–20, Institute for Marine Dynamics. 16 Spencer, D., 1992, “A standard method for the conduct and analysis of ice resistance model tests,” in Proceedings of the 23rd American Towing Tank Conference, New Orleans.

140 / Vol. 127, MAY 2005

Transactions of the ASME

Downloaded 22 May 2008 to Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm