Ocean Engineering 37 (2010) 14641476

Contents lists available at ScienceDirect

Ocean Engineering
journal homepage: www.elsevier.com/locate/oceaneng

Squat prediction in muddy navigation areas

G. Delefortrie a,n, M. Vantorre b, K. Eloot a,b, J. Verwilligen a, E. Lataire b
a b

Flanders Hydraulics Research, Berchemlei 115, 2140 Antwerp, Belgium Ghent University, Maritime Technology Division, Technologiepark 904, 9052 Ghent, Belgium

a r t i c l e in f o
Article history: Received 5 May 2010 Accepted 22 August 2010 Editor-in-Chief: A.I. Incecik Available online 15 September 2010 Keywords: Mud Sinkage Trim Mathematical model

a b s t r a c t
Common squat prediction formulae to assess the navigation safety usually do not take into account the bottom condition. Nevertheless, the presence of a uid mud layer is not an uncommon condition in conned areas where accurate squat predictions are necessary. From 2001 to 2004 an extensive experimental research program was carried out to measure the manoeuvring behaviour of deep drafted vessels in muddy areas. A part of the program focused on the undulations of the watermud interface and their relationship to the ships squat. Mostly the sinkage of the ship is damped due to the presence of the mud layer, but a larger trim can occur due to the watermud interface undulations. This article presents a mathematical model to predict the squat in muddy navigation areas. & 2010 Elsevier Ltd. All rights reserved.

1. Introduction Squat, dened as the sinkage and trim of vessels due to their own forward speed, is of particular importance in shallow water areas. Small under keel clearances cause large return currents which lead to important sinkages and higher risks of bottom touching as already mentioned by Constantine (1960). In shallow navigation areas the presence of a soft uid mud layer on the bottom is not exceptional, but its effect is mostly neglected in the formulation of squat. As a consequence pilots and scientists may disagree on the safety of navigation. Mostly pilots have to rely on the high frequency echo to determine the water depth. As the latter detects the top of the mud layer and not the solid (or nautical) bottom level, they may still be able to navigate safely through a muddy navigation area, even in case the ship is navigating at a zero (or even negative) under keel clearance according to the echo sounder. In cases where common squat formulae would predict grounding of a ship navigating in a shallow fairway at a rather high speed, the presence of a mud layer may prevent such grounding. Indeed a limited or even negative under keel clearance referred to a mud layer does not necessarily lead to impracticable manoeuvres as mentioned by Delefortrie et al. (2007). When initially the ship has a small under keel clearance referred to the mud layer, she may hit the mud layer due to squat. This mud,

having a larger density than water, will affect the buoyancy of the ship and will probably smoothen the squat effect. However, to be sure about the mud effect additional research had to be carried out, because the literature offers very limited results on this topic.

2. State of the art 2.1. General research on squat Scientic research on squat took off with Constantine (1960) who discussed the different squat behaviour for subcritical, critical and supercritical vessel speeds. In the subcritical domain (Frh o1) Tuck (1966) proved that for open water conditions of constant depth the sinkage and trim of the vessel to be linear with the parameter
rh gFrh q 2 1Frh

F2

In which Frh represents the depth related Froude number

2 Frh

V2 gh

Corresponding author. Tel.: + 32 3 224 69 62; fax: + 32 3 224 60 36. E-mail addresses: Guillaume.Delefortrie@mow.vlaanderen.be (G. Delefortrie), Marc.Vantorre@UGent.be (M. Vantorre), Katrien.Eloot@mow.vlaanderen.be (K. Eloot), Jeroen.Verwilligen@mow.vlaanderen.be (J. Verwilligen), Evert.Lataire@UGent.be (E. Lataire). 0029-8018/$ - see front matter & 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.oceaneng.2010.08.003

This theory was later extended to dredged channels by Beck et al. (1975). Naghdi and Rubin (1984) offer some reections on Tucks theory and introduce a new one. An analogous theory has been developed by Cong and Hsiung (1991). Ankudinov and Daggett (1996) however are pessimistic about the complexity of numerical theories. For this reason, several

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1465

Nomenclature AEP AR ai B b bi c CB ci CS CT D d di DP E e ei f f0 Frh g gi h h* i0 i1 ji ki LPP P p0 q0 expanded area ratio of propeller () rudder area (m2) regression coefcient, (i0,1) () ship beam (m) mud type, Table 2 () regression coefcient, (i 0,1,2) () mud type, Table 2 () block coefcient () regression coefcient, (i0,1,2) () dimensionless sinkage, Eq. (4) () dimensionless trim, Eq. (5) () 6000 TEU container ship model () mud type, Table 2 () regression coefcient, (i0,r) () propeller diameter () tanker model () mud type, Table 2 () regression coefcient, (i0,1) () mud type, Table 2 () regression coefcient () depth related Froude number, Eq. (2) () mud type, Table 2 () regression coefcient, (i0,h) () total depth (m) mud type, Table 2 () hydrodynamically equivalent depth (m) regression coefcient () regression coefcient () regression coefcient, (i0, r) () regression coefcient, (i0, r) () ship length (m) propeller pitch (m) regression coefcient () regression coefcient ()

S si T TEU TP U ukc V VT x zA zF

a g
z zMAX

m mcrit r r*
P Ph F Fij 0 Fij F00 ij F000 00

solid bottom condition () regression coefcient (i0,n) () ship draft (m) (number of) twenty feet equivalent unit containers () thrust (N) 8000 TEU container ship model () under keel clearance () ship speed (m/s) propeller induced speed, Eq. (22) (m/s) longitudinal coordinate, positive towards the stern (m) sinkage aft (m) sinkage fore (m) increase parameter, Eq. (31) () parameter, Eq. (1) () amplitude of rising (m) maximal amplitude of rising (m) dynamic viscosity (Pa s) critical dynamic viscosity (Pa s) density (kg/m3) dimensionless density, Eq. (16) ( ) keel penetration parameter, Eq. (21) () keel penetration parameter, Eq. (14) () uidization parameter () regression coefcient (i0,h) (j 0,r) () regression coefcient (i0,h) (j 0,r) () regression coefcient (i0,h) (j 0,r) () regression coefcient ()

Subscripts n 1 2 propeller related to water related to mud

authors carried out experimental research as Dand (1972) and Gourlay (2000) who offers a solution for squat prediction with random bottom conditions to endorse their theories. Jiang and Henn (2003) present a numerical method valid from subcritical to supercritical speed. An overview of slender body methods is given in Gourlay (2008). More practical methods based on experimental research are presented by Barrass (1979), however his results could not be validated by Seren et al. (1983). Barrass (2004) gives an overview of the work he performed on squat. More general overviews are given by Dumas (1982); Blaauw and Van der Knaap (1983), Millward (1990) and PIANC (1997) working group 30. Interesting full scale measurements were carried out by Ankudinov et al. (2000), Stocks et al. (2004), Harting et al. (2009) and Harting and Reinking (2002) among others. Most discussions focus on ships sailing in open water or in rectangular shaped canals without drift angle or propulsion. In some cases the drift angle was considered, as by Von Bovet (1985), Martin and Puls (1986), de Koning Gans and Boonstra (2007) and Eloot et al. (2008).

The oldest results are presented by Sellmeijer and van Oortmerssen (1983) who also registered undulations in the water mud interface. The sinkage is less above mud in comparison with the solid bottom condition and decreases with increasing layer thickness. The mud density does not seem to have any effect. Vantorre and Coen (1988) showed that three speed ranges can be detected for the behaviour of the water mud interface:

 At low speed a small sinkage near the fore body is detected,

which disappears amidships and turns into an elevation abaft;

 At a certain speed value the sinkage at the entrance changes 

suddenly into an elevation. The section at which the jump occurs moves abaft with increasing speed; If the speed increases more, the rising of the interface occurs behind the stern. The amplitude of the elevation can exceed the mud layer thickness several times.

2.2. Research on squat in muddy areas The research on squat in muddy areas is a topic that has not been tackled thoroughly. Only three research institutes carried out experimental research focussing on the hydrodynamic forces.

The latter occurs at a speed, which for inviscid uids is given by the theoretical expression s   8 r 3 Vcrit gh1 1 1 1m1 3 r2 27 m1 being the blockage of the ship in the water layer, meaning the ratio of the ships immersed cross sectional area and the canals cross section. Subscript 1 refers to the water layer, subscript 2 to the mud layer.

1466

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

The sinkage of the vessel is related to these speed ranges, but experimental results do not always follow the theory. At low speeds the mud layer causes a very slight increase of sinkage while at higher speed a sinkage decrease with mud layer is observed together with an increase of trim. Brossard et al. (1990) described that the sinkage is identical as in the solid bottom condition when the ships keel does not penetrate the mud layer. An effect is observed at negative under keel clearances:

2004 at the Towing Tank for Manoeuvres in Shallow Water cooperation Flanders Hydraulics Research, Ghent University. The formulae are valid within the range of conditions covered by this experimental research. The shallow water towing tank (88 m 7 m 0.5 m) is equipped with a planar motion carriage, a wave generator and an auxiliary carriage for shipship interaction tests. Thanks to computerized control and data-acquisition, the facilities are operated in a fully automated manner. The carriage runs 24/7 without the need for permanent surveillance. 3.2. Ship models Most runs in muddy navigation areas were carried out with (Table 1) a 6000 TEU container carrier (D), which was the design ship for the Belgian harbours at that time. The mathematical model will be based on the measurements carried out with this ship. Additional runs in a selection of conditions were carried out with a scale model of a 8000 TEU container carrier (U) and a scale model of a tanker (E) to assess the inuence of the hull form, see 4.4. 3.3. Bottom conditions Mud was simulated by a mixture of two types of chlorinated parafn and petrol, so that both density and viscosity could be controlled within certain ranges. For environmental reasons, the tank was divided into three compartments: a test section, a mud reservoir and a water reservoir. Bottom and tank walls were protected with a polyethylene coating. No viscosity or density gradients were included. The selected bottom conditions are represented in Table 2. The densityviscosity combinations were based on measurements of density and rheology proles carried out in the outer harbour of Zeebrugge, Belgium in 19971998. A mud layer conguration is dened by two characters: a letter (b,y,h) denoting the material characteristics and Figs. 13 re-presenting the layer thickness. Tests carried out above a solid bottom are referred to as S. The gross under keel clearance relative to the tank bottom was varied between 7% and 32% of draft, yielding 12% to +21% ukc relative to the mudwater interface. Throughout this paper, the interface watermud will be used as a reference for expressing the under keel clearance, unless specied otherwise. 3.4. Test types Stationary captive motion model tests were carried out in each combination of mud layer and realistic under keel clearance at different speeds, from 2 kn ahead to 10 kn ahead at least in steps of 2 kn (full scale values).

 The rigidity of the mud has only a small effect;  The density gradient signicantly affects the sinkage: the 
higher the gradient, the smaller the sinkage. It is assumed that the buoyancy is an important factor; Adding currents leads to further reduction of the sinkage.

The trim of the vessel is only signicantly affected by rigid mud. In this case an increase of trim with increasing density gradient was observed. The sign of the trim changes when penetrating the mud. Doctors et al. (1996) showed that for the ship hydrodynamics a shallow water approach can serve as a quite reliable approximation for analyzing the case of a viscous lower layer, where the mud viscosity can be interpreted as an effective reduction in the total depth of the water. In spite of these results, until now no sufciently correct models have been presented to predict the squat in muddy areas.

3. Experimental setup 3.1. Test facilities The new squat formulae presented in this article are all derived from experimental research carried out from 2001 until
Table 1 Ship models (even keel). Model Scale LPP (m) B (m) T (m) CB AR (m2) # blades DP (m) P/DP () AEP () D 1/75 289.8 40.25 13.50 0.59 60.96 5 8.145 0.97 0.8 U 1/80 331.8 42.8 14.54 0.65 83.13 6 8.46 1.00 0.96 E 1/75 286.8 46.8 15.5 0.82 98.34 5 7.73 0.65 0.62

Table 2 Bottom conditions and tested models. Mud type Density (kg/m3) Dynamic viscosity (Pa s) Layer thickness 0.75 m 1 d c b f h e g fresh water sea water S 1100 1150 1180 1200 1210 1260 1250 1000 1025 solid bottom 0.03 0.06 0.10 0.11 0.19 0.29 0.46 0.0010 0.0012 D/E D D D/E D/E 1.50 m 2 D/E D D D D/E D D/E 3.00 m 3 D/E/U D D D D/E

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1467

The propeller rate was varied at 6 kn ahead between 60% and 100% of the maximal propeller rate. Some runs were carried out with a drift or rudder angle or with astern speeds, but these runs will not be discussed in this article. 3.5. Measurements The sinkage of the ship was measured at four positions on the hull: starboard side fore and aft and portside fore and aft.

A positive trim angle is measured when the sinkage at the bow is larger than at the stern. As the literature already mentioned the occurrence of undulations of the water mud interface measures were taken to register them. A device which follows the level of the mud layer (mufo) and one which follows the water level (wafo) are assembled on a

1.5

1
+ h2

0.5

h2

() x LPP () 1

dy = 0.28 B dy = 1.30 B dy = 2.32 B

0 -1 -0.5 0 0.5 1.5 2

Fig. 1. Positioning the tripods in the towing tank.

Fig. 3. Undulations of the interface at various lateral distances of the ship D. Mud f2, + 3.9% ukc, Frh 0.38, no propeller action. The ship is represented taking squat into account and sails to the left.

Fig. 2. Arrangement of the wave meters in the towing tank.

1468

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

tripod (see Fig. 1). In total three tripods were assembled in order to register the undulations of the interface. The mufo consists of a oater, which is resistant to the articial mud. The density of the oater is situated between the water and the mud density. The position of the oater is therefore similar to the amplitude of the interface. The oater is attached to a disk, which reects a laser beam. The variations of the laser beam are measured 20 times per second and register the actual position of the mud layer. The wafo is based upon the principles of the potentiometer. A constant electric current is sent through a string which has a homogeneous resistance. The voltage is therefore also constant. A second electrode is a tube made of stainless steel. When the water level in the tube changes a proportional change of voltage will be measured. Three tripods were placed in the towing tank as close as possible to the passing ship. The lateral distance between the ship and the mufos is as shown in Fig. 2. Depending on the ships velocity the position of the interface was measured 1020 times per second. The tripods were placed in the middle of the tank so that the position of the interface could be measured before, while and after the ship was passing.

The rising increases with increasing speed, but this increase is limited once the undulations are behind the ship, see Fig. 5. This is especially the case with low density mud layers (c, d). When the vessel navigates above the mud layer the rising will increase faster with the velocity when the density and viscosity of the mud layer are closer to water. With thinner mud layers the rising becomes only signicant once the viscosity drops below a certain critical value, which is located between 0.12 and 0.18 Pa s. A signicant undulation is always observed when the ship navigates in contact with the mud layer, see Fig. 6. The rising is mostly located amidships for higher density mud layers. For lower density and viscosity the rising is located abaft, as shown in Fig. 7. The transitory situation is a rising occurring in two phases.

4. Observations 4.1. Undulations of the water mud interface Fig. 3 gives an example of the measured undulations of the water mud interface, which seem to behave as a Kelvin pattern. The maximal amplitude closest to the ship is represented in Fig. 4.
2 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1 0 0.1 mud G 0.2 mud H Frh (-) mud B 0.3 mud C 0.4 mud D 0.5

2 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1 0 0.1 mud F 0.2 mud H Frh (-) mud B 0.3 mud C 0.4 mud D 0.5

Fig. 6. Maximal rising in function of ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 1.1% under keel clearance referred to the water mud interface.

MAX

+ h2 () h2

1 0.75 x () LPP 0.5 0.25 0 -0.25 0 0.1 mud F mud D 0.2 Frh (-) mud H amidship mud B aft mud C 0.3 0.4 0.5

+ h2
MAX

() h2

Fig. 4. Maximal rising in function of ship speed. Ship D. Thickness of the mud layer: 3 m full scale. No propeller or rudder action. 9.8% under keel clearance referred to the water mud interface.

Fig. 7. Longitudinal position, at which the rising is maximal, in function of ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 1.1% under keel clearance referred to the water mud interface.

2.2
()

() + h2
MAX

0.75 0.5 0.25 0 0 0.1 mud G mud D 0.2 Frh (-) mud H amidship mud B aft mud C 0.3 0.4 0.5

2 1.8

x LPP

h2

1.6 1.4 1.2 1 -100% -50% mud G 0% 50% mud D 100%

propeller rate (-) mud H mud B mud C

Fig. 5. Longitudinal position, at which the rising is maximal, in function of ship speed. Ship D. Thickness of the mud layer: 3 m full scale. No propeller or rudder action. 9.8% under keel clearance referred to the water mud interface.

Fig. 8. Maximal rising in function of the propeller rate. Ship D. Thickness of the mud layer: 3 m full scale. No rudder action. Ship speed: 6 kn full scale. 12% under keel clearance referred to the water mud interface.

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1469

0.75 0.5 0.25 0 -0.25 -100%

x LPP

()

-50% mud G mud D

0% propeller rate (-) mud H aft mud B amidship

50% mud C

100%

Fig. 9. Longitudinal position, at which the rising is maximal, in function of the propeller rate. Ship D. Thickness of the mud layer: 3 m full scale. No rudder action. Ship speed: 6 kn full scale. 12.2% under keel clearance referred to the water mud interface.

0.16 0.14 0.12 CS (-) 0.08 0.06 0.04 0.02 0 0 0.1 solid mud B 0.2 Frh (-) mud F mud C mud G mud D mud H 0.3 0.4 0.5 0.1

0.06 0.05 0.04 CS (-) 0.03 0.02 0.01 0 -100% -50% mud G 0% 50% propeller rate (-) mud H mud B mud C mud D 100%

Fig. 10. Sinkage in function of the ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 3.9% under keel clearance referred to the water mud interface.

Fig. 12. Sinkage in function of the propeller rate. Ship D. Thickness of the mud layer: 3 m full scale. No rudder action. 12.2% under keel clearance referred to the water mud interface.

4.2. Sinkage
0.16 0.14 0.12 CS (-) 0.1 0.08 0.06 0.04 0.02 0 0 0.1 solid mud E 0.2 Frh (-) mud F mud C mud G mud D mud H 0.3 0.4 0.5

The mean sinkage as measured during the tests will be represented dimensionless as CS Cs 100 zF zA 2LPP 4

The values zA and zF are positive downwards. Figs. 10 and 11 give an overview of the ships sinkage in function of the speed for different bottom conditions. The following can be observed:

 When the ship navigates above the mud layer the rising of the
interface is signicantly larger for mud layers with a viscosity below a critical viscosity. When the under keel clearance is small, this can eventually result in contact between the vessel and the mud layer. The mud will yield an increase of buoyancy, which results in a decrease of the sinkage; If the ships keel penetrates the mud layer, the large rising amidships, which occurs for higher density mud layers, will cause an increase of buoyancy. The sinkage will consequently be smaller.

Fig. 11. Sinkage in function of the ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 1.1% under keel clearance referred to the water mud interface.

Reversed propeller action in case of navigating ahead yields a relatively large rising near the propeller, see Figs. 8 and 9. In this case the pattern of the undulations is rather random. The rising does not seem to start abaft the ship in the given experimental speed range, although some experimental speeds were higher than those predicted by Eq. (3). A possible explanation is the higher viscosity of the mud layers in this experimental program and the assumption of inviscid uids in Eq. (3).

The sinkage, for a same small under keel clearance referred to the solid bottom, is always larger above a solid bottom than above a muddy bottom. The same observations were made by Sellmeijer and Van Oortmerssen (1983) and Vantorre and Coen (1988), nevertheless the latter mentioned a slight increase of sinkage at low speeds.

1470

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

A working propeller generates an additional longitudinal speed which changes the pressure balance and thus the squat of the ship. As shown in Fig. 12 the additional sinkage is more or less quadratic with the propeller rate. 4.3. Trim As the sinkage is not constant along the ships hull, the ship will be dynamically trimmed. For slender hulls this generally results in a larger sinkage at the stern, while full body ships mostly have a larger sinkage at the bow. When the ship navigates in a muddy area, the trim will be inuenced as well and its absolute value will usually increase, due to the extra asymmetry in the buoyancy caused by the rising of the interface. The dimensionless total trim CT as measured during the tests CT 100 zF zA LPP 5

 In all cases a larger rising causes a larger asymmetry and thus a 

larger trim. This is in accordance with the observations made by Vantorre and Coen (1988). A change of trim sign when penetrating the mud as reported by Brossard et al. (1990) is not observed.

As for the sinkage propeller action inuences the trimming of the ship. Propeller action yields a larger dynamic trim, especially with propeller action astern, see Fig. 15. This coincides with the effect of propeller action on the rising of the water mud interface as shown in Fig. 9.

is represented for different navigation conditions in Figs. 13 and 14, where a negative trim means a larger sinkage abaft. It can be stated that, in combination with the observations of the undulations of the interface:

 A rising will have the largest inuence on the trim when it 

takes place amidships. The inuence will decrease when the rising moves abaft; The trim will be smaller when the top of the rising is wider;
0.005 0 -0.005 CT (-) -0.01

0 -0.005 -0.01 -0.015 -0.02 -0.025 -0.03 -0.035 -0.04 -0.045 -0.05 -100%

CT (-)

-50% mud G

0% propeller rate (-) mud B

50% mud C mud D

100%

mud H

Fig. 15. Trim as a function of the propeller rate. Ship D. Thickness of the mud layer: 3 m full scale. No rudder action. 12.2% under keel clearance referred to the water mud interface.

2 () + h2
MAX

-0.015 -0.02 -0.025 0 0.1 solid mud B 0.2 Frh (-) mud F mud C mud G mud D mud H 0.3 0.4 0.5

1.8 1.6

h2

1.4 1.2 1 0 0.1 U, -12.2% ukc U, +10% ukc 0.2 0.3 Frh (-) D, -12.2% ukc E, -9.4% ukc 0.4 D, +10% ukc E, -4.4% ukc 0.5

Fig. 13. Trim in function of the ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 3.9% under keel clearance referred to the water mud interface.

Fig. 16. Maximal rising in function of ship speed and ship type. Thickness of the mud layer D: 0.04 m model scale. No propeller or rudder action.

0.005 0 -0.005 CT (-) -0.01 -0.02 -0.025 -0.03 -0.035 0 0.1 solid mud E 0.2 Frh (-) mud F mud C mud G mud D mud H 0.3 0.4 0.5
CS (-) 0.25 0.2 0.15 0.1 0.05 0 0 0.1 U, -12.2% ukc U, +10% ukc
Fig. 14. Trim in function of the ship speed. Ship D. Thickness of the mud layer: 1.5 m full scale. No propeller or rudder action. 1.1% under keel clearance referred to the water mud interface.

-0.015

0.2 Frh (-)

0.3

0.4 D, +10% ukc E, -4.4% ukc

0.5

D, -12.2% ukc E, -9.4% ukc

Fig. 17. Sinkage in function of the ship speed and ship type. Thickness of the mud layer D: 0.04 m model scale. No propeller or rudder action.

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1471

4.4. Effect of the hull form Fig. 16 shows the maximal rising of the interface for the three ships above or in contact with the thickest mud layer D. For both container carriers the trend is more or less the same, but for the tanker the rising seems to reach a maximum at low speed. For the fuller ship the rising tends to occur fore or amidships, an increase of speed will cause a shift towards the stern of the ship, but not necessarily an increase of amplitude. For a same Froude number the dimensionless sinkage will be larger for the fuller ship, while both container carriers follow the same trend, see Fig. 17. The same can be concluded for the trim, Fig. 18, but the fuller ship has a larger bow sinkage.

0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 0

CS = 0.5942 R2 = 0.9826

CT,S (-)

CT = -0.1919 R2 = 0.9893

0.05

0.1

0.15

0.2

0.25

(-)
Fig. 19. Application of expressions (6) and (7). Ship D. Thickness of mud layer G: 3 m full scale. No propeller or rudder action. 12.2% under keel clearance referred to the water mud interface.

5. Mathematical model 5.1. Denitions A mathematical model will be built to predict the ships squat when sailing straight ahead based on the measurements carried out with the 6000 TEU container carrier. The effect of a rotating propeller will be taken into account. According to Tuck (1966) the mean sinkage of the vessel in the subcritical speed range can be modelled as Cs s0 gFrh 6

With s0 a coefcient to be derived from regression analysis. An analogous relationship is valid for the vessels trim CT t0 gFrh 7

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -0.1 -0.2 0 2

s0 Linear trend Quadratic trend

s0, t0 (-)

t0

6 T/(h-T) (-)

10

12

This relationship is also valid in muddy navigation areas, even when the ship is penetrating a highly viscous mud layer, as shown in Fig. 19. Eqs. (6) and (7) can consequently be used to predict the squat in muddy areas. However some physical awareness is needed. Tuck (1966) developed his expressions for an open shallow water environment without any mud layer. The critical speed, which is related to the return current, will certainly be affected by the presence of a mud layer. Nonetheless Eqs. (6) and (7) will be used as a starting point to develop new expressions. 5.2. Effect of the water depth Instead of trying to nd a new expression for the subcritical speed in function of the muddy environmental conditions the coefcients s0 and t0 will be formulated as mud dependent
0.08 0.06 0.04 CT (-) 0.02 0 -0.02 -0.04 -0.06 0 0.1 U, -12.2% ukc U, +10% ukc 0.2 Frh (-) D, -12.2% ukc E, -9.4% ukc D, +10% ukc E, -4.4% ukc 0.3 0.4 0.5

Fig. 20. Regression coefcients s0 (6) and t0 (7) in function of the under keel clearance above a solid bottom. Ship D. No propeller or rudder action.

parameters. Furthermore even above a solid bottom some discrepancies occur in function of the under keel clearance, e.g. Fig. 20. To take account of very shallow water effects Eqs. (6) and (7) should thus be reformulated as   T 8 a0 gFrh CS a1 hT " Ct b2  T hT 2 # T b1 b0 gFrh hT

5.3. The hydrodynamically equivalent depth The effect of the presence of the mud layer can be modelled with Eqs. (8) and (9) using a hydrodynamically equivalent depth h* instead of the real depth h. With h2 the thickness of the mud layer and h1 the height of the upper lying water layer, the total depth can be written as h h1 h2 10

The bottom material can vary from water over soft mud to consolidated mud. If the mud has large viscosity and density values, like sand or clay, the material will hardly move when a ship passes by and its top can be considered as the actual seabed. In this case the hydrodynamically equivalent depth h* is h* h1 11

Fig. 18. Trim in function of the ship speed and ship type. Thickness of the mud layer D: 0.04 m model scale. No propeller or rudder action.

1472

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

On the other hand, if the material is very uid the mud layer cannot be considered as a solid bottom. In the limit condition of two equivalent water layers, the hydrodynamically equivalent depth is h* h1 h2 h 12
s0 (-)

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

Solid

h1

h2

h3

For intermediate situations a parameter F can be dened, so that h* h1 Fh2 rh 13

Particular values for the parameter F are 0 ( hard layer of thickness h2) and 1 ( watery layer of thickness h2), F represents consequently the degree of watery behaviour of the bottom layer and is therefore called the uidization parameter. Intuitively the uidization parameter of the mud covering the seabed depends on the following aspects:

6 8 T/(h*-T) (-)

10

12

14

Fig. 22. Regression coefcients s0 (6) in function of the hydrodynamic equivalent under keel clearance above mud h. Ship D. No propeller or rudder action.

 the rheological properties (e.g. viscosity) of the mud: a 

decrease of the latter means a more uid mud layer and will logically result in an increased uidization parameter; the under keel clearance referred to the mudwater interface: the uidization parameter increases when the ships keel is located closer to the mud or penetrates the mud. In these conditions the mud layer is stirred and will behave more uidly.

2.5 2 1.5 (-) 1 0.5 0 -0.5

3m

1.5 m

0.75 m

The assumption the mud layer does not affect the critical speed means that the parameter g will always be expressed with h instead of h*. 5.4. Effect of the mud layer 5.4.1. Without propulsion The effect of the mud layer on the hydrodynamic force balance could be modelled using the uidization parameter. Figs. 21 and 22 show the effect of using the same concept on the regression coefcient s0. A reasonable agreement can be observed. The hydrodynamic equivalent under keel clearances shown in Figs. 21 and 22 have been determined with a uidization parameter that takes into account the position of the ships keel referred to the top of the mud, expressed by Th1 Ph h1 14

-1 -0.2

-0.15

-0.1

-0.05

0 h (-)

0.05

0.1

0.15

0.2

Fig. 23. Fluidization parameter to determine the hydrodynamic equivalent under keel clearance for sinkage prediction above mud layers of low viscosity. Ship D. No propeller or rudder action.

2.5 2 1.5 (-) 1 0.5 0 -0.5 -1 -0.2

3m

1.5 m

0.75 m

and the composition of the mud. The following interpretations can be derived from Figs. 23 and 24:

 When penetrating the mud the uidization parameter is

signicantly larger than 1, meaning that for an equal ships
Solid b1 b2 b3

-0.15

-0.1

-0.05

0.8 0.7 0.6 0.5 s0 (-) 0.4 0.3 0.2 0.1 0 0

0 h (-)

0.05

0.1

0.15

0.2

Fig. 24. Fluidization parameter to determine the hydrodynamic equivalent under keel clearance for sinkage prediction above mud layers of high viscosity. Ship D. No propeller or rudder action.


2 4 6 T/(h*-T) (-) 8 10 12

Fig. 21. Regression coefcients s0 (6) in function of the hydrodynamic equivalent under keel clearance above mud b. Ship D. No propeller or rudder action.

speed the sinkage will be smaller due to the presence of the mud. This can be related to the changed buoyancy, nevertheless a signicant density effect cannot be observed. This is in accordance with Sellmeijer and van Oortmerssen (1983). On the other hand, if the keel does not penetrate the mud layer, the sinkage will become larger compared to a solid bottom and this for a same total depth. This is rather relative as the sinkage will always be smaller for larger under keel clearances above any bottom condition. The decrease of the uidization parameter with increasing under keel clearance referred to the top of the mud layer is

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1473

more or less equal for all mud compositions. On the other hand, the decrease will start at a larger under keel clearance for mud layers having a low viscosity, which can be ascribed to the buoyancy effect of the higher risings in these conditions. The effect of the muds viscosity, as mentioned in 4.1, has thus a signicant inuence on the sinkage of the ship. The uidization parameter can consequently be written as

Also a new hydrodynamically equivalent depth is needed to assess the inuence of the thrust in muddy navigation areas. For mud layers having a high viscosity the uidization parameter dening h* in Eq. (24) can be written simply as

F f0

26

If the viscosity drops below the critical one, the uidization parameter changes to

F c1 r*Ph c2 m c0
With r* the dimensionless density

15

F gh Ph g0

27

r*

r2 r1 r1

This uidization parameter does not increase further once 16

Ph Z0.05. Figs. 26 and 27 show the low density mud D as an

example. The total trim taking account of propeller action can be written accordingly CT t0 gV,h tn gVT ,h 28

In Eq. (15) the following restrictions have to be taken into account:

Ph Z 0:15
Ph c2 r 0

17 18 19

m 4 mcrit : c2 0

Eq. (17) states the uidization parameter will not decrease innitely with increasing under keel clearance, while Eq. (18) allows a constant uidization parameter when the keel (almost) penetrates the mud. The parameter c2 takes into account the higher undulations for mud layers having a viscosity below the critical one. An analogous expression can be built for the trim of the vessel. In this case the uidization parameter can be written as   T T 20 F d0 dr r*P F00 Fh0 r* F0r Fhr h2 h2 which uses an alternative formulation for the position of the ships keel referred to the mud layer

Fig. 25 shows the effect of the under keel clearance above a solid bottom linearly expressed   T CT VT ,h* i1 29 i0 gVT ,h h*T As for the sinkage a new hydrodynamically equivalent depth is needed to express the propeller effect in muddy areas   h h 30 F j0 jr r*P F000 F0h0 2 r* F00r F0hr 2 T T However this is only valid for propeller action ahead. Fig. 15 showed a larger longitudinal speed VT is needed to predict the
0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 -0.02 -0.04 -0.06 -0.08 0 Linear trend

Th1 h2

21
sn, tn (-)

No restrictions apply to Eq. (20). Both Eqs. (15) and (20) are valid within the experimental scope. 5.4.2. With propulsion A working propeller generates an additional longitudinal speed VT that can be estimated as a function of the thrust generated by the propeller v  u u 1:5TP   t 22 VT signTP  rpD2  P This additional speed changes the pressure distribution along the ships hull and thus the squat. The factor 1.5 in Eq. (22) has been determined experimentally for an under keel clearance of 26% above a solid bottom so that the total sinkage in this condition can be written as Cs s0 gV,h gVT ,h CS V,h CS VT ,h 23

6 T/(h-T) (-)

10

12

Fig. 25. Regression coefcients sn (25) and tn (28) in function of the under keel clearance above a solid bottom. Ship D. Effect of propeller action.

0.14 0.12 0.1 sn (-) 0.08 0.06 0.04 0.02 0 0

Solid

d1

d2

d3

The effect of the under keel clearance on the thrust dependent term needs however a new set of regression coefcients ei, see Fig. 25   T CS VT ,h* e1 e0 gVT ,h h*T 24 sn gVT ,h The total sinkage, including propeller action is then CS s0 gV,h sn gVT ,h 25

10

15

20

25

T/(h*-T) (-)
Fig. 26. Regression coefcients sn (25) in function of the hydrodynamic equivalent under keel clearance above mud D. Ship D. Effect of propeller action.

1474

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1.2 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -0.2

d1

d2

d3

0.18 0.16 0.14 CS model (-) 0.12 0.1 0.08 0.06 0.04 0.02 0 -0.05

R2 = 0.9502

(-)

-0.15

-0.1

-0.05

0.05

0.1

0.15

0.2

h (-)
Fig. 27. Fluidization parameter to determine the hydrodynamic equivalent under keel clearance for sinkage prediction above the light mud layer D. Ship D. Effect of propeller action.

0.05 0.1 CS measurement (-)

0.15

0.2

Fig. 30. Comparison between measured and modelled sinkage values: all runs. Ship D.

0
-1.6

-0.01 CT model (-) -0.02 -0.03 -0.04 -0.05

R2 = 0.9289

-1.8

(-)

-2

-2.2

-0.06 -0.06
-2.4 0 2 4 6 T/(h-T) (-) 8 10 12

-0.05

-0.04

-0.03 -0.02 CT measurement (-)

-0.01

0.01

Fig. 31. Comparison between measured and modelled trim values: all runs. Ship D.

Fig. 28. Increase of propeller induced longitudinal speed due to astern rotation. Ship D. Inuence of the under keel clearance above a solid bottom.

measurement 0.18 0.16 0.14 0.12 CS (-) 0.1 0.08 0.06 0.04 0.02 0 0 0.1

model

0 -0.01 CT (VT) (-) -0.02 -0.03 -0.04 -0.05 -0.06 0 0.2 26% ukc 0.4 0.6 (VT) (-) 0.8 32% ukc 1 1.2

Increasing propeller rate 100% -60%

0.2 Frh (-)

0.3

0.4

0.5

15% ukc 10% ukc

Linear trend

Fig. 29. Evaluation of formula (29). Solid bottom condition, inuence of propeller action, both ahead and astern. Ship D.

Fig. 32. Comparison between measured and modelled sinkage values. 10% under keel clearance above a solid bottom. Ship D. Runs with propeller action are labelled.

effect of propeller action astern on the trim. For trim Eq. (22) should be replaced by s   1:5TP   31 VT aUsignTP  rpD2  With a 1 for positive values of the propellers thrust. For an astern rotating propeller the increase a depends on the under keel clearance, see Fig. 28 TP o0 : a k1 T k0 h*T 32

When Eq. (32) is used to correct VT, Eq. (29) can be used to predict the trim for any propeller rate above a solid bottom, see Fig. 29. The correction coefcient a depends also on the composition of the mud layer. The hydrodynamically equivalent depth in Eq. (32) is a function of a uidization parameter. For mud layers below the critical viscosity this parameter can be determined as

F0 p0 P F0000

h2 00 F h0 r*F00hr T

33

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

1475

0 -0.01 -0.02 CT (-) -0.03 -0.04 -0.05

measurement

model 60% 100%

6. Conclusions and recommendations For a same small under keel clearance referred to the solid bottom the sinkage will be mostly smaller when a mud layer is present. This is not always the case for the ships trim. The squat of the ship can be related to the observed undulations of the water mud interface, that become dominant once the viscosity of the mud layer is below a critical one. A mathematical model predicting fairly well the ships squat for container carriers has been built taking into account the bottom conditions and propeller action. Therefore, the principle of a hydrodynamically equivalent depth has been used. For several uidization parameters a different formulation is needed depending on the viscous characteristics of the mud layers. This can be linked to a critical viscosity as observed with the undulations of the water mud interface. As this critical viscosity lies somewhere between 0.11 and 0.19 Pa s a linear interpolation between the formulae should be applied within this viscosity range. Future efforts will be undertaken to include more parameters (drift angle,y) in the mathematical model and to investigate how the mud layer affects the critical speed regimes.

-60% -0.06 0 0.1 0.2 Frh (-)

Fig. 33. Comparison between measured and modelled trim values. 10% under keel clearance above a solid bottom. Ship D. Runs with propeller action are labelled.

0.3

0.4

0.5

0.14 0.12 0.1 CS (-) 0.08 0.06 0.04 0.02 0 0

measurement

model

Acknowledgements
Increasing propeller rate 100% -60%

0.1

0.2 Frh (-)

0.3

0.4

0.5

Fig. 34. Comparison between measured and modelled sinkage values. Ship D. Thickness of the mud layer c: 3 m full scale. 7.2% under keel clearance referred to the water mud interface. Runs with propeller action are labelled.

The data presented in this article were obtained during the research project Determination of the nautical bottom in the harbour of Zeebrugge: Nautical implications, which was carried out co-operatively by Ghent University and Flanders Hydraulics, commissioned by T.V. Noordzee & Kust (Ostend, Belgium) in the frame of the optimisation of the maintenance dredging contract for the harbour of Zeebrugge, nanced by the Department Maritime Access, a division of the Mobility and Public Works department of the Flemish Government. References
Ankudinov, V., Daggett, L., 1996. Squat predictions for manoeuvring applications. In: Proceedings of MARSIM 1996 (International Conference on Marine Simulation and Manoeuvrability), Copenhagen, Denmark, pp. 467495. Ankudinov, V., Daggett, L., Hewlett, J., Jakobsen, B. 2000. Prototype measurement of ship sinkage in conned water. In: Proceedings of MARSIM 2000 (International Conference on Marine Simulation and Manoeuvrability), Orlando, USA, p 233. Barras, C., 1979. The phenomena of ship squat. International Shipbuilding Progress 26, 4447. Barras, C. 2004. Thirty-two years of research into ship squat. Squat workshop 2004, Elseth, Germany, pp. 125. Beck, R., Newman, J., Tuck, E., 1975. Hydrodynamic forces on ships in dredged channels. Journal of Ship Research 19 (3), 166171. Blaauw, H., Van der Knaap, F., 1983. Prediction of squat of ships sailing in restricted water. In: Proceedings of the International Harbour Congress, pp. 8193. Brossard, C., Delouis, A., Galichon, P., Granboulan, J., Monadier, P., 1990. Navigability in channels subject to siltation. In: Proceedings of the 22nd International Coastal Engineering Conference, Delft, The Netherlands, pp. 30883101. Cong, L., Hsiung, C., 1991. Computing wave resistance, wave prole and sinkage and trim of transom stern ships. Marine and Offshore Operations, 99112. Constantine, T., 1960. On the movement of ships in restricted waterways. Journal of Fluid Mechanics 9 (2), 247257. Dand, I., 1972. On full form ships in shallow water: some methods for the prediction of squat in subcritical ows. National Physical Laboratory, Ship Division, Ship Report 160. de Koning Gans, H., Boonstra, H., 2007. Squat effects of very large container ships with drift in a harbor environment. In: Proceedings of the International Maritime-port Technology and Development Conference, Singapore, pp. 613620. Delefortrie, G., Vantorre, M., Verzhbitskaya, E., Seynaeve, K., 2007. Evaluation of safety of navigation in muddy areas through real time manoeuvring simulation. Journal of Waterway, Port, Coastal and Ocean Engineering 133 (2), 125135. Doctors, L.J., Zilman, G., Miloh, T., 1996. The inuence of a bottom mud layer on the steady-state hydrodynamics of marine vehicles. In: Proceedings of the 21st Symposium on Naval Hydrodynamics, pp. 727 742.  Dumas, R., 1982. Le surenfoncement des navires du a leur mouvement. Revue Technique du Service des Phares et Balises 53, 2658 In French. Eloot, K., Verwilligen, J., Vantorre, M., 2008. An overview of squat measurements for container ships in restricted water. In: Varyani, K.S. (Ed.), International

measurement 0 -0.005 -0.01 -0.015 CT (-) -0.02 -0.025 -0.03 -0.035 -0.04 -0.045 0 0.1

model

60% 100%

-60% 0.2 Frh (-) 0.3 0.4 0.5

Fig. 35. Comparison between measured and modelled trim values. Ship D. Thickness of the mud layer c: 3 m full scale. 7.2% under keel clearance referred to the water mud interface. Runs with propeller action are labelled.

While more viscous mud layers can use a simpler expression

F q0 P F00000
5.5. Results

34

The overall comparison between measured and modelled values is shown in Figs. 30 and 31. Some outliers occur, but the overall accuracy is fairly well. Some spot checks in different conditions are shown in Figs. 3235.

1476

G. Delefortrie et al. / Ocean Engineering 37 (2010) 14641476

Conference on Safety and Operations in Canals and Waterways SOCW 2008, 1516 September 2008 Glasgow, UK, pp. 106116. Gourlay, T., 2000. Mathematical and computational techniques for predicting the squat of ships. Ph.D. thesis, University of Adelaide, Department of Applied Mathematics. Gourlay, T., 2008. Slender-body methods for predicting ship squat. Ocean Engineering 35 (2), 191200. Harting, A., Reinking, J., 2002. SHIPS: a new method for efcient full-scale ship squat determination. In: Proceedings of the PIANC 30th International Navigation Congress, Sydney, Australia, pp. 8491. Harting, A., Laupichler, A., Reinking, J., 2009. Considerations on the squat of unevenly trimmed ships. Ocean Engineering 36, 193201. Jiang, T., Henn, R., 2003. Numerical prediction of ships squat and trim in shallow water. In: Proceedings of MARSIM 2003 (International Conference on Marine Simulation and Manoeuvrability), Kanazawa, Japan. Martin, H., Puls, D., 1986. Vertical forces, trim moments and changes of draught and trim of ships in shallow water. Schiffbauforschung 25 (3), 155159. Millward, A., 1990. A preliminary design method for the prediction of squat in shallow water. Marine Technology 27 (1), 1019.

Naghdi, P., Rubin, M., 1984. On the squat of a ship. Journal of Ship Research 28 (2), 107117. PIANC, 1997. Approach channels A guide for design. Final report of the joint Working Group PIANC and IAPH, in cooperation with IMPA and IALA. Supplement to PIANC Bulletin, No. 95, 108 p. Sellmeijer, R., van Oortmerssen, G., 1983. The effect of mud on tanker manoeuvres. The Royal Institution of Naval Architects, Spring Meetings 1983, paper no. 7. Seren, D., Ferguson, A., McGregor, R., 1983. Squatan examination of two practical prediction methods. The Naval Architect 5, 228230. Stocks, D., Daggett, L., Page, Y., 2004. Maximization of ship draft in the St.Lawrence Seaway. Squat workshop 2004, Elseth, Germany. Tuck, E., 1966. Shallow water ow past slender bodies. Journal of Fluid Mechanics 26 (1), 8195. Vantorre, M., Coen, I., 1988. On sinkage and trim of vessels navigating above a mud layer. The Royal Society of Flemish Engineers, International Harbour Congress. Von Bovet, G., 1985. Tiefgangs- un Trimmwinkelanderungen von Schiffen under Berucksichtigung endlicher Wassertiefe un Einub auf die Steuerbarkeit. Schiffbauforschung 24 (1), 311 (in German).