Sie sind auf Seite 1von 48

NUMERICAL STUDY ON TRIM OPTIMIZATION FOR

180,000 CBM Class Twin LNG Carrier (2963/2964)

HYUNDAI
KNUTSEN

July 2017
HYUNDAI
KNUTSEN
i

ABSTRACT

A study on optimum trim of the 180,000 CBM Class Twin Skeg LNG Carrier was conducted
using HMRI’s own HiFoam code, developed on the basis of OpenFOAM. The resistance of a
ship at model scale was calculated with various drafts, speeds and trims. In addition the effective
and delivered powers at full scale were predicted using ITTC 78 method and experimental data
in HMRI.
The calculation conditions were determined as follows.

Draft Trims
9.25 m -1.5 -1.0 -0.5 0.0 +0.5 (5 trims)
11.50 m -1.5 -1.0 -0.5 0.0 +0.5 (5 trims)
Speed (knots) : 15.0 18.5 19.5 20.5 22.5 (5 speeds)
* (+) Trim by Head; (-) Trim by Stern

HYUNDAI
KNUTSEN

Revision Summary
No. Date Contents

HMRI-2017-AB-103
ii

CONTENTS
Page
ABSTRACT ............................................................................................................ i
CONTENTS ............................................................................................................ ii
LIST OF TABLES ................................................................................................... iii
LIST OF FIGURES ................................................................................................. iv

1. INTRODUCTION ............................................................................................... 1
2.COMPUTATIONAL CONDITIONS ................................................................... 2
3. RESULTS ............................................................................................................ 3
4. REFERENCES .................................................................................................... 4

APPENDIX Ⅰ NOMENCLATURE...................................................................................................
12
APPENDIX Ⅱ NUMERICAL METHODS .......................................................................................
33
APPENDIX Ⅲ TEST PROCEDURE AND ANALYSIS METHOD .................................................
39

HYUNDAI
KNUTSEN

HMRI-2017-AB-103
iii

LIST OF TABLES
Page
Table 1 Principal particulars of ship ..............................................................................................
5
Table 2 Variation of ship particulars with computational conditions ..............................................
5
Table 3 Computed resistance of model at Teq=9.25m (Unit: N) ........................................................
6
Table 4 Computed resistance of model at Teq=11.50m (Unit: N) ......................................................
6
Table 5 Predicted effective power at Teq=9.25m (Unit: kW) .............................................................
7
Table 6 Predicted effective power at Teq=11.50m (Unit: kW) ...........................................................
7
Table 7 Predicted delivered power at Teq=9.25m (Unit: kW) ............................................................
8
Table 8 Predicted delivered power at Teq=11.50m (Unit: kW) ..........................................................
8

HYUNDAI
KNUTSEN

HMRI-2017-AB-103
iv

LIST OF FIGURES
Page
Figure 1 Computational domain and grid arrangement for the numerical simulations .............................
9
Figure 2 Delivered power with various speed and trim conditions at 9.25m ............................................
10
Figure 3 Delivered power with various speed and trim conditions at 11.5m.............................................
11
Figure 4 Wave around FP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 9.25m condition).....................................................................
12
Figure 5 Wave around AP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 9.25m condition).....................................................................
13
Figure 6 Wave around FP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 9.25m condition).....................................................................
14
Figure 7 Wave around AP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 9.25m condition).....................................................................
15
Figure 8 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 9.25m condition).....................................................................
16
Figure 9 Wave around AP with various speeds under constant trim and draft conditions;

Figure 10
HYUNDAI
(0.5m trim by stern with draft of 9.25m condition).....................................................................
Wave around FP with various speeds under constant trim and draft conditions;
17

Figure 11 KNUTSEN
(Even trim with draft of 9.25m condition) ..................................................................................
Wave around AP with various speeds under constant trim and draft conditions;
18

(Even trim with draft of 9.25m condition) ..................................................................................


19
Figure 12 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 9.25m condition) .....................................................................
20
Figure 13 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 9.25m condition) .....................................................................
21
Figure 14 Wave around FP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 11.50m condition) ...................................................................
22
Figure 15 Wave around AP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 11.50m condition) ...................................................................
23
Figure 16 Wave around FP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 11.50m condition) ...................................................................
24
Figure 17 Wave around AP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 10.50m condition)...................................................................
25
Figure 18 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 11.50m condition) ...................................................................
26

HMRI-2017-AB-103
v

Figure 19 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 11.50m condition) ...................................................................
27
Figure 20 Wave around FP with various speeds under constant trim and draft conditions;
(Even trim with draft of 11.50m condition) ................................................................................
28
Figure 21 Wave around AP with various speeds under constant trim and draft conditions;
(Even trim with draft of 11.50m condition) ................................................................................
29
Figure 22 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 11.50m condition) ...................................................................
30
Figure 23 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 11.50m condition) ...................................................................
31

HYUNDAI
KNUTSEN

HMRI-2017-AB-103
1

1. INTRODUCTION

As computational technology develops, CFD plays an important role in maritime industries.


Recently, it has been revealed that the resistance of a model ship (RTM) calculated from CFD
simulation is highly correlated with the break horse power estimated from towing tank
experiment in the previous study [1]. CFD simulations have been carried out on the model ship
for 180,000 CBM Class Twin Skeg LNG Carrier (2963/2964) which has a scale of 1/34.8000
to obtain RTM with various draft, speed and trim conditions. Principal particulars of the
objective ship are shown in Table 1.

The conditions for simulation are given as following.

Draft Trims
9.25 m -1.5 -1.0 -0.5 0.0 +0.5 (5 trims)
11.50 m -1.5 -1.0 -0.5 0.0 +0.5 (5 trims)
Speed (knots) : 15.0 18.5 19.5 20.5 22.5 (5 speeds)
* (+) Trim by Head; (-) Trim by Stern

HYUNDAI
The HiFoam created by HMRI based on the OpenFOAM is applied to the present
simulation. Its performance on accuracy was validated in the previous study [2] by comparing
the results with commercial code (STAR-CCM+) and it was found that the code provided

KNUTSEN
satisfying or even better results occasionally than the commercial code.

The estimated resistance of model ship from CFD is used to predict effective power and
delivered power with the aid of ITTC 78 method.

HMRI-2017-AB-103
2

2. COMPUTATIONAL CONDITIONS

- Computational Grid

The flow field around the objective ship should be divided into a number of small cells
having finite volume to carry out numerical simulations. In the present simulation, trimmed
mesh provided by STAR-CCM+ was applied to construct a grid system and the resultant grid
system consisted of about 1.3 million cells for resistance simulations and about 1.8 million cells
for self-propulsion simulations. The computational domain and grid arrangements around the
objective ship are shown in Figure 1 and Figure 2. The half of ship model was considered only
to reduce the computational effort and the symmetric plane was applied to reflect the rest. The
computational domain was in the shape of rectangular with overall length of 3.5 times to Lpp of
the model ship and overall height of 2.5 times to Lpp. The Neumann condition is applied to side
of computational domain.

To resolve the flows near the wall, the ship was covered by the prism layers where the y+
value of the first points lies within the range of 80~100. As boundary conditions for turbulent

HYUNDAI
flows on the wall, standard wall functions were applied for turbulent kinetic energy and
turbulent viscosity while the modified wall function was applied for turbulent dissipation rate.

- Physical Models
KNUTSEN
As mentioned in the preceding section, HiFoam was applied in order to carry out the
numerical simulation of flows around the ship. The volume of fluid (VoF) method was
employed to resolve the free surface of the water and the realizable k-epsilon model was used as
the turbulence model in simulations for predicting resistance. The Reynolds Stress Model
(RSM) with linear pressure strain model was used to achieve higher accuracy in self-propulsion
simulations since the wake at propeller plane acts an important role in predicting the
performance of the propeller. Fresh water at 15°C where its properties can be found in the ITTC
Recommended Procedure was considered for computation. The code also considered the ship
motion such as trim and sinkage by applying DFBI (Dynamic Fluid Body Interaction). The 2nd-
order convection schemes were selected for spatial discretization and the local time stepping
method was used for temporal discretization. Detailed information of numerical method is
explained in Appendix II.

HMRI-2017-AB-103
3

3. RESULTS

Tables 3 and 6 show the computed RTM of model ship and the predicted effective power.
Tables 7 to 8 shows the predicted delivered power and variations of the delivered power are
shown in Figures 3 to 4. The processes for predicting the effective power and delivered power
are explained in Appendix III in detail. Note that the values presented in the figures are ratios of
delivered power at trimmed condition to even condition under the same speed and draft
conditions. The figures measure relative delivered powers with respect to that of even condition
only. Therefore it is unlikely to say that lower delivered power ratio with higher speed under the
same trim and draft condition guarantees lower delivered power.

Figures 5 to 23 show wave patterns obtained from the CFD simulations. As long as trim and
draft remain constant, the wave length near AP increases and waves after transom become more
complicated as the speed of ship increases. The wave in the near downstream of the transom
tends to rise higher with increase of draft under the same trim and speed conditions.

HYUNDAI
KNUTSEN

HMRI-2017-AB-103
4

4. REFERENCES
1. Park, S. H., “Estimation of the Optimum Trims in Container Carriers by using Numerical
Prediction of Ship Resistances,” Journal of The Society of Naval Architects of Korea
(submitted), 2014
2. Lee, S. B., “Application of OpenFOAM to Prediction of Hull Resistance,” 8th International
OpenFOAM Workshop, 11-14 June, 2013, Jeju, Korea
3. “Model Tests for 180,000 CBM Class Twin LNG Carrier (2963/2964)”, HMRI Report,
HMRI-2017-AB-089, 2017

HYUNDAI
KNUTSEN

HMRI-2017-AB-103
5

Table 1 Principal particulars of ship


Model : T834
Scale : 34.8000
:LOA LBP B D
(m) (m) (m) (m)
299.00 293.60 48.00 26.40

Table 2 Variation of ship particulars with computational conditions

Draft Trim* L WL  S LCB


(m) (m) (m) (m3) (m2) (m)


-1.5 292.3 92444.0 15110.0 -5.738
-1.0 291.4 92444.0 15101.0 -4.738
9.25
HYUNDAI
-0.5
0.0
+0.5
290.4
289.3
288.1
92444.0
92444.0
92444.0
15083.0
15076.0
15069.0
-3.747
-2.773
-1.810

11.50
KNUTSEN
-1.5
-1.0
-0.5
296.5
296.2
295.7
117696.0
117696.0
117696.0
16755.0
16739.0
16718.0
-6.653
-5.790
-4.940
0.0 295.2 117696.0 16699.0 -4.100
+0.5 294.6 117696.0 16677.0 -3.271
*
(+) for Trim By Head, (-) for Trim By Stern

(+) for forward, (-) for afterward

HMRI-2017-AB-103
6

Table 3 Computed resistance of model at Teq=9.25m (Unit: N)

Trim -1.5 -1.0 -0.5 0.0 m +0.5


Speed (Trim by Stern) (Trim by Stern) (Trim by Stern) (Even keel) (Trim by Head)

15.0 knots 36.52 36.44 36.28 36.22 36.19


18.5 knots 52.81 52.74 52.65 52.69 52.77
19.5 knots 58.35 58.47 58.54 58.67 58.85
20.5 knots 65.08 65.11 65.21 65.50 65.79
22.5 knots 81.21 81.32 81.46 82.02 82.43

HYUNDAI
Table 4 Computed resistance of model at Teq=11.50m (Unit: N)

Speed
Trim
KNUTSEN
-1.5
(Trim by Stern)
-1.0
(Trim by Stern)
-0.5
(Trim by Stern)
0.0 m
(Even keel)
+0.5
(Trim by Head)

15.0 knots 37.84 37.86 37.88 37.92 37.95


18.5 knots 56.07 56.09 56.12 56.17 56.23
19.5 knots 62.33 62.32 62.35 62.39 62.46
20.5 knots 69.21 69.20 69.18 69.24 69.37
22.5 knots 86.33 86.32 86.30 86.26 86.48

HMRI-2017-AB-103
7

Table 5 Predicted effective power at Teq=9.25m (Unit: kW)

Trim -1.5 -1.0 -0.5 0.0 m +0.5


Speed (Trim by Stern) (Trim by Stern) (Trim by Stern) (Even keel) (Trim by Head)

15.0 knots 7010 6994 6964 6952 6946


18.5 knots 12611 12595 12574 12583 12602
19.5 knots 14864 14894 14912 14945 14991
20.5 knots 17715 17723 17750 17829 17908
22.5 knots 25277 25313 25355 25530 25657

HYUNDAI
Table 6 Predicted effective power at Teq=11.50m (Unit: kW)

Speed
Trim KNUTSEN
-1.5
(Trim by Stern)
-1.0
(Trim by Stern)
-0.5
(Trim by Stern)
0.0 m
(Even keel)
+0.5
(Trim by Head)

15.0 knots 7259 7263 7267 7274 7279


18.5 knots 13559 13563 13570 13582 13596
19.5 knots 16043 16038 16047 16057 16076
20.5 knots 18948 18945 18940 18956 18991
22.5 knots 26927 26921 26916 26904 26974

HMRI-2017-AB-103
8

Table 7 Predicted delivered power at Teq=9.25m (Unit: kW)

Trim -1.5 -1.0 -0.5 0.0 m +0.5


Speed (Trim by Stern) (Trim by Stern) (Trim by Stern) (Even keel) (Trim by Head)

15.0 knots 8527 8453 8372 8289 8214


18.5 knots 15355 15272 15152 15044 14978
19.5 knots 18203 18070 17982 17905 17844
20.5 knots 21817 21696 21597 21573 21549
22.5 knots 31252 31131 31000 31065 31071

HYUNDAI
Table 8 Predicted delivered power at Teq=11.50m (Unit: kW)

Speed
Trim

15.0 knots
KNUTSEN
-1.5
(Trim by Stern)

8405
-1.0
(Trim by Stern)

8353
-0.5
(Trim by Stern)

8319
0.0 m
(Even keel)

8303
+0.5
(Trim by Head)

8289
18.5 knots 15827 15732 15668 15638 15613
19.5 knots 18724 18615 18538 18508 18485
20.5 knots 22114 21998 21931 21881 21870
22.5 knots 31457 31321 31205 31175 31192

HMRI-2017-AB-103
9

1 Lpp

FLOW

HYUNDAI
WATER
SURFACE
1.5 Lpp

KNUTSEN
SHIP

3.5 Lpp
1.5 L
pp

Figure 1 Computational domain and grid arrangement for the numerical simulations

HMRI-2017-AB-103
10

HYUNDAI
KNUTSEN

Figure 2 Delivered power with various speed and trim conditions at 9.25m
Trim(m) : (+) Head, (-) Stern

HMRI-2017-AB-103
11

HYUNDAI
KNUTSEN

Figure 3 Delivered power with various speed and trim conditions at 11.5m
Trim(m) : (+) Head, (-) Stern

HMRI-2017-AB-103
12

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 4 Wave around FP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
13

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 5 Wave around AP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
14

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 6 Wave around FP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
15

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 7 Wave around AP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
16

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 8 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
17

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 9 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 9.25m condition)

HMRI-2017-AB-103
18

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 10 Wave around FP with various speeds under constant trim and draft conditions;
(Even trim with draft of 9.25m condition)

HMRI-2017-AB-103
19

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 11 Wave around AP with various speeds under constant trim and draft conditions;
(Even trim with draft of 9.25m condition)

HMRI-2017-AB-103
20

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 12 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 9.25m condition)

HMRI-2017-AB-103
21

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 13 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 9.25m condition)

HMRI-2017-AB-103
22

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 14 Wave around FP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 11.50m condition)

HMRI-2017-AB-103
23

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 15 Wave around AP with various speeds under constant trim and draft conditions;
(1.5m trim by stern with draft of 11.50m condition)

HMRI-2017-AB-103
24

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 16 Wave around FP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 11.50m condition)

HMRI-2017-AB-103
25

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 17 Wave around AP with various speeds under constant trim and draft conditions;
(1.0m trim by stern with draft of 10.50m condition)

HMRI-2017-AB-103
26

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 18 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 11.50m condition)

HMRI-2017-AB-103
27

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 19 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by stern with draft of 11.50m condition)

HMRI-2017-AB-103
28

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 20 Wave around FP with various speeds under constant trim and draft conditions;
(Even trim with draft of 11.50m condition)

HMRI-2017-AB-103
29

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 21 Wave around AP with various speeds under constant trim and draft conditions;
(Even trim with draft of 11.50m condition)

HMRI-2017-AB-103
30

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 22 Wave around FP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 11.50m condition)

HMRI-2017-AB-103
31

(a)
15.0
knots

(b)
18.5
knots

(c)
19.5
knots
HYUNDAI
KNUTSEN
(d)
20.5
knots

(e)
22.5
knots

Figure 23 Wave around AP with various speeds under constant trim and draft conditions;
(0.5m trim by head with draft of 11.50m condition)

HMRI-2017-AB-103
32

APPENDIX Ⅰ
NOMENCLATURE

SYMBOL DESCRIPTION UNIT

AT transverse area above waterline m2

B breadth moulded at mid-ship m

D depth m

F flux passing through a surface of a cell m3 s

k turbulent kinetic energy m2 s 2

LCB longitudinal center of buoyancy m

L PP length between perpendiculars m

L WL length at waterline m

Pk production of the turbulent kinetic energy kg m  s3

RTM

S HYUNDAI
total resistance of a model
wetted surface area
N

m2

Sbk

TA KNUTSEN
wetted surface area of bilge keels

draft at after perpendicular


m2

TF draft at fore perpendicular m

TM draft at mid-ship m

Teq equivalent draft m

U velocity vector of a fluid m s

 displacement m3

 volume fraction of a fluid in a cell -

 turbulent dissipation rate m2 s 3

 dynamic viscosity kg m  s

 kinematic viscosity m2 s

 density kg m3

HMRI-2017-AB-103
33

APPENDIX Ⅱ

NUMERICAL METHODS

Flow and related phenomena can be described by partial differential equations which, in
general, cannot be solved analytically. To obtain an approximate solution numerically, we have
to use a discretization method which approximates the differential equations by a system of
algebraic equations, which can then be solved on a computer. The approximations are applied to
small domains in space and/or time so the numerical solution provides results at discrete
locations in space and time.

1. Discretization using Finite Volume Method

A flow field of interest should be subdivided into numerous small domains as shown in figure
A.1 to produce a computational mesh on which the governing equations are subsequently solved.
The discretization procedure converts the governing equations in partial differential form into a
set of algebraic equations that are commonly expressed in matrix form as following.
[A]{x}  {b} (A.1)

source vector. HYUNDAI


where [A] is a square matrix, {x} is the column vector of dependent variable and {b} is the

Finite Volume discretization of each term is formulated by first integrating the term over a

KNUTSEN
cell volume V. Most spatial derivative terms are then converted to integrals over the cell surface
S bounding the volume using Gauss’s theorem.
    dV   dS   (A.2)
V S

where S is the surface area vector,  can represent any tensor field and the  is used to
represent any tensor product, i.e. inner, outer and cross and the respective derivatives:
divergence    , gradient  and    . Volume and surface integrals are then linearized
using appropriate schemes.

HMRI-2017-AB-103
34

Figure A.1 Control volume

The Laplacian Term

The Laplacian term is integrated over a control volume and linearized as follows:
      dV   dS        f S f    f
V S f
(A.3)

The face gradient discretization is implicit when the length vector d between the center of the
cell of interest P and the center of a neighboring cell N is orthogonal to the face plane.
 N  P
S f     f  S f (A.4)
d
In the case of non-orthogonal meshes, an additional explicit term is introduced which is
evaluated by interpolating cell center gradients, themselves calculated by central differencing
cell center values.

The Convection Term

The convection term is integrated over a control volume and linearized as follows:
     U  dV   dS    U    S f    U  f  f   F f (A.5)

HYUNDAI
V S f f

where F is the surface flux. The face field  f can be evaluated using a variety of schemes such
as central differencing (CD), upwind differencing (UD) and blended differencing schemes

KNUTSEN
which combine UD and CD in an attempt to preserve boundedness with reasonable accuracy.

The Temporal Term

The first time derivative  t is integrated over a control volume as follows:



t V
 dV (A.6)

The term is discretized by simple differencing in time using new values  n at the time step
we are solving for and old values  o that were stored from the previous time step. When the
Euler implicit scheme is applied, the Eq. (A.6) can be discretized as follows:
   V     PPV 
n o


t V
 dV  P P
t
(A.7)

2. Modeling of Governing Equations

A behavior of non-Newtonian fluids can be expressed by using the Navier-Stokes equations


but the general and analytic solutions of the equations are not known yet. Alternatively we can
predict the flow fields by solving the approximated Navier-Stokes equations but it is even
difficult to solve the approximated equations directly for complex flows such as turbulent flows.

HMRI-2017-AB-103
35

Thus we need to model the real flows to make it possible to handle with computers.

Mean Flow Field

Flows of incompressible and viscous fluids can be resolved by solving the incompressible
Navier-Stokes equations:
U  0 (A.8)
U
    UU     U   p (A.9)
t

The above equations are the continuity and momentum equations, respectively. The left-hand-
side terms of the Eq. (A.9) are temporal, convection and diffusion terms, respectively. Since it is
not easy to solve the non-linear systems, it is preferable to linearize the convection term of the
momentum equation which is non-linear by using the Eq. (A.5). In this case the flux F is the
mass flow rate passing through a face of a cell and it should satisfy the continuity equation.
By applying the discretization methods mentioned in the previous section, the momentum
equation can be expressed as a set of linear equations as follows.
aP UnP   aN UnN  R P (A.10)

HYUNDAI
N

where a is the coefficients of the matrix [A] of Eq. (A.1).


One can obtain the velocity field by solving the Eq. (A.10) but the pressure is not updated yet.

KNUTSEN
Furthermore the resultant velocity field does not satisfy the continuity condition. For the
incompressible Navier-Stokes equations, it is common to resolve such problems by dealing with
the pressure-velocity coupling such as SIMPLE algorithm.
In order to derive the pressure equation, a semi-discretized form of the momentum equation is
used:
aP UnP  H  U   p (A.11)

The H  U  term consists of two parts: the transport part including the matrix coefficients for
all neighbors multiplied by corresponding velocities and the source part including the source
part of the transient term and all source parts but the pressure gradient. From the above equation,
we can get an expression for U at the center of a cell P as follows.
H U 1
UP   p (A.12)
aP aP
Once again we can estimate the velocities at the surfaces by interpolating the above
expression.

 H U   1 
Uf   
 aP   aP 
p   f
(A.13)
 f f

where a bar over a variable denotes that the quantity is obtained by interpolating values at the
centers of cells adjacent to the face.

HMRI-2017-AB-103
36

From the continuity condition,   U  0 and we can obtain the pressure equation from the
continuity condition and Eq. (A.12) as follows.
 1   H U 
   p       (A.14)
 aP   aP 
Finally the above equation can be expressed in the discrete form as follows.
 1    H U 
 S   a   p  f    S 



(A.15)
f
  P f
 f
 a P f

The face flux F is then calculated using the Eq. (A.13).


 H U    
    1 p
F  S  U f  S    
  aP   aP  f
  
f 
(A.16)
   f 
Solution procedure for the Navier-Stokes system using PISO algorithm is shown in figure A.2.
The main difference between PISO and SIMPLE algorithm is existence of the second pressure
correction equation.

START

HYUNDAI Set up the initial conditions for all field values

KNUTSEN
Solve the momentum equation

Solve the pressure correction equation

Correct pressure and velocities

Solve the second pressure correction equation

Correct pressure and velocities

Solve all other discretised transport equations

NO
Converged?

YES

END

Figure A.2 Flow chart of PISO algorithm

Volume of Fluid Method

Volume of Fluid (VoF) method is a CFD technique to simulate multiphase flow simulations
including free surfaces which was presented by Hirt and Nichols and relies on the definition of
an indicator function. This function indicates that a cell is occupied by a fluid or another, or a

HMRI-2017-AB-103
37

mix of both. By introducing the indicator function, the governing equations for the conventional
VoF method can be written as follows.
U  0 (A.17)

    U   0 (A.18)
t

  U
     UU   p    T   fb (A.19)
t
where U represents the velocity field shared by the two fluids throughout the flow domain,
 is the phase fraction, T  2S  2    U  I 3 is the deviatoric viscous stress tensor with
the mean rate of strain tensor S  0.5  U   U   and I   ij ,  is density, p is
T
 
pressure and fb are body forces per unit mass. In VoF simulations the latter forces include
gravity and surface tension effects at the interface. The phase fraction can take values within the
range 0    1 . The flow properties can be calculated as weighted averages as follows.
  l   g 1    (A.20)

  l   g 1    (A.21)

where subscripts l and g denote the liquid and gaseous phases of the fluids, respectively. At

HYUNDAI
the interface of the liquid and gas fluids the surface tension is modeled using the continuum
surface force model as follows.
f   (A.22)

KNUTSEN
where  is the mean curvature of the free surface. And it is common to define a modified
pressure in order to simplify the definition of boundary conditions as follows.
pd  p   g  x (A.23)

where x is the position vector.

Turbulence Model

Reynolds averaged Navier-Stokes equation (RANS) is the one of the most widely used
method to simulate the turbulent flows. Among the RANS models, two equation models such as
k-epsilon or k-omega models are the most popular for the engineering practices. In the present
report the realizable k-epsilon model is employed which is a variant of a standard k-epsilon
model satisfying certain mathematical constraints on the normal stresses, consistent with the
physics of turbulent flows. The transport equations of the turbulent kinetic energy and the
dissipation rate of the realizable k-epsilon model can be written as follows.
  k 
     Uk      ek   Pk  Pb    YM  Sk (A.24)
t

    2 
     U      e    C1S   C2  C1 C3 Pb  S (A.25)
t k   k

HMRI-2017-AB-103
38

where
   k
C1  max 0.43, S S  2tr SS
   5  
In the above equations, Pk represents the production of turbulence kinetic energy due to the
mean velocity gradients and Pb is the production of turbulent kinetic energy due to buoyancy.
YM represents the contribution of the fluctuating dilatation in compressible turbulence to the
overall dissipation rate which can be omitted for the incompressible flows. The turbulent
viscosity is modeled as follows.
k2
t   C  (A.26)

where
1
C  U *  tr SS  tr WW
A0  As kU * 

W  W  2 ijk k W  W   ijk k

where W is the mean rotation rate tensor viewed in a rotating reference frame with the
angular velocity k . The model constants A0 and As are given by

HYUNDAI
A0  4.04 As  6 cos 

 
1 Sij S jk Ski
  cos1 S  tr SS
3
KNUTSEN
6W

The model coefficients are given as follows.


W
S3

C1  1.44 C2  1.9  k  1.0    1.2

HMRI-2017-AB-103
39

APPENDIX Ⅲ

TEST PROCEDURE AND ANALYSIS METHOD

1. Effective power prediction

Method
The ship resistance in model scale is determined by CFD calculation at various drafts, trims
and speeds. The CFD resistance values are corrected according to the experimental resistance
values of the identical or equivalent class hull form. During a normal resistance calculation,
rudder(s) and bilge keel(s) are not included

Analysis
 ITTC1978 3-Dimemsional Method
The conversion from model(M) to ship(S) is made according to the 1978 ITTC Performance
Prediction method. This implies that the frictional resistance( C F ) is calculated from the ITTC
1957 model-ship correlation line, giving the relation between C F and Reynolds number(Rn)
0.075 VL V
CF  , Rn  , Fn 
(log 10Rn  2) 2
 gL

HYUNDAI
where g = acceleration of gravity, in m/s2
L = Length of waterplane for ship and model respectively, in m

KNUTSEN
V = speed, in m/s
 = kinematic viscosity, in m2/s (ITTC 1960)

It is further assumed that the form factor(k) based upon ITTC 1957 and the residuary
resistance coefficient( C R ) are identical for model and ship at the same Fn. The total ship
resistance coefficient( CTS ) can be calculated as follows.
SS  SBK
CTS  [(1  k )CFS  CF ]  CR  CAA ,
SS
R TM
CR  CTM  (1  k )CFM , CTM 
MSM VM
1 2
2

where C AA = air-resistance coefficient


CTM and CTS = the total model and ship resistance coefficient
R TM = model total resistance, in N
SS , SBK = wetted surface area of ship and bilge keel respectively, in m2
VM = model speed, in m/s
C F = hull roughness allowance coefficient
 M = mass density of tank water, in Ns2/m4

The form factor(k) is determined on the basis of the Prohaska’s proposal with some

HMRI-2017-AB-103
40

modification using resistance values in the low speed range(Froude number range of 0.1 to 0.2).
It seems most satisfactory to evaluate the exponent of Fn in order to obtain the best
approximation of measured data points. Normally this would results in :
CTM Fn n
 (1  k )  c  ( )
CFM CFM
where n, c and k are to be determined by a least square method.
A power of Fn =4  n  6 is more appropriate for full ships.
In case that the resistance data at the low speed are not enough to determine the form factor,
the form factor is determined on the basis of the existing experimental resistance results and
assumed that it is identical.
The hull roughness allowance coefficients( C F ) is assumed to be:

k S 1/ 3
CF  [105( )  0.64]  103 with hull roughness k S =150  m
L

The air resistance coefficients( C AA ) is assumed to be:

HYUNDAI
AT
C AA  0.001
SS
where A T = transverse projected area above the water of the ship including superstructures, in
m2
KNUTSEN
The total ship resistance can be calculated from :
1
R TS  CTS  SSS VS
2

Finally, the effective power( PE ) of ship is calculated from:

PE  R TS  VS

2. Delivered power prediction

Method
The trimmed delivered power is predicted from the existing experimental results of the
identical or equivalent class hull form without self-propulsion calculation. Within HMRI’s
model test data base, the ratio of the delivered power between even and trimmed condition is
assumed to have a linear relation to the ratio of the effective power with some correlation factor
according to the draft condition.

DHPTRIM EHPTRIM
= ∙ω
DHPEVEN EHPEVEN

where 𝜔 = correlation factor

HMRI-2017-AB-103
HYUNDAI
KNUTSEN

Hyundai Maritime Research Institute (HMRI)

Das könnte Ihnen auch gefallen