Conference of Global Chinese Scholars on Hydrodynamics

HYDRODYNAMICS IN DEEPWATER TLP TENDON DESIGN

WANG Tao, ZOU Jun Houston Offshore Engineering, 200 West Lake Park Blvd, Suite 1300, Houston, TX 77079, USA E-mail : [email protected]

ABSTRACT: This paper introduces hydrodynamic aspects of in-place TLP tendon design and analysis with emphases on TLP hull/tendon/riser coupled dynamic analysis; tendon Mathieu instability; tendon springing and ringing responses; tendon bottom tension slacking due to wave overtopping; and tendon vortex induced vibration (VIV). KEY WORDS: TLP, hull/tendon/riser, coupled dynamic analysis, Mathieu instability, springing and ringing, tension slacking, wave overtopping, VIV

tendon transverse motions. It is meaningful for robust design by checking Mathieus equation to ensure no instability problems. It is well known that both tendon springing and tendon ringing are high-frequency tendon responses due to high-frequency resonant motions. However, ringing appears to be extremely bursting and transient (Natvig, 1994). Extensive experimental studies and
H UL L P O NT O O N E X T E NS IO N

1. Introduction TLP tendons are pre-tensioned slender members, their tops are connected to the hull and bottoms are moored to the seabed, as illustrated in Fig.1. In deepwater, the TLP hull tends to interact more pronouncedly to its tendons and risers in terms of mass, stiffness and damping coupling. The TLP hull/tendon/riser coupled dynamic analysis forms a consistent analysis and design framework for deepwater field development and application. A traditional way of simulating the dynamics of a TLP is to use an uncoupled method, which ignores all or part of the coupling effects (mass, stiffness and damping) between the TLP hull and tendons/risers. Ormberg and Larsen (1998) found that the uncoupled analysis produced severely inaccurate results, especially for ultra deepwater application. Coupled time domain analysis technologies (Paulling and Webster, 1986; Kim et al., 1994; Ma et al., 2000; Zou, 2003) have been established. Further developments and validations have been reported by Zou (2003). Mathieus equation was applied in the investigation of the parametric responses excited by the interactions of tendon tension variation and lateral dynamics of tendons (Zhang et al., 2002). Mathieus instability could be triggered at a certain model of tendon lateral motion dependent on the magnitude of fluctuation of the tendon tension and damping of

T O P C O N N E C TO R A S S Y

TEN D ON POR C H

L E N G T H A D JU S T M E N T JO I N T

LO A D M E A S U R E M E NT UN IT 3 2 O D T O 4 0 O D T R A N S ITIO N T E N D O N P IP E

P IP E TO P IP E W E L D A R E A

4 0 O D T O 3 2 O D T R A N S IT IO N

T E N D O N P IP E P IP E T O P IP E W E L D

E X TE N S IO N

D R IV E H E A D A N D G U I D E C O N E

T E ND O N B O T TO M RE C E P T A C L E

T E N D O N P IL E A N C H O R

Fig. 1 TLP tendon components

Biography: WANG Tao, Senior Naval Architect; ZOU Jun, Manager of Naval Architecture

numerical simulations (Zou and Kim, 1996; Zou, 1997;

386

Zou et al., 1998) have been undertaken to investigate

the tendon springing and ringing responses. The results reveal that the springing is due to weak asymmetric waves while the ringing is due to strong asymmetric waves. As observed in the North Sea, strong nonlinear waves in the irregular wave train do create the ringing. Tendon springing responses have significant impact on tendon fatigue while tendon ringing responses have considerable impact on the tendon extreme strength. In 2005, two strong hurricanes, Katrina and Rita, hit Gulf of Mexico (GOM), causing tremendous damages and even capsizing one mini-TLP. It brought great interest in the investigation of tendon bottom tension slacking due to wave overtopping on the top of column(s) and/or even on top of the deck. Prior to hurricane Katrina and Rita, we had investigated the wave overtopping effects on tendon dynamic responses on a three-column TLP. In 1998, physical wave basin model tests were performed and numerical simulations by model the model (Zou et al., 2003) were completed and compared with the measured results. The results indicated that the wave overtopping on top of column was the source to cause tendon tension slacking. It has been reported (Leverette et al., 2003) that the excitation of VIV on tendons due to GOMs warm core eddies and deep submerged jets (also called cold eddies) can generate noticeable responses of the whole TLP platform. Observations of these excitations due to VIV are typically very high-frequency responses, higher than those of heave, roll/pitch natural frequencies. This might imply the hull structural and/or deck structural modes have been excited by tendon VIV. The impacts of tendon VIV are 1) reducing tendon fatigue life, 2) inducing excessive operating downtime, 3) affecting operating personnel comforts, 4) creating hull and deck structural high-frequency fatigue. The outline of this paper is as follows. First, the hull/tendon/riser coupled dynamic analysis was briefly introduced. Second, the progress of tendon Mathieus instability was reviewed. Third, tendon springing and ringing were discussed. Fourth, tendon bottom tension slacking due to wave overtopping was highlighted. Fifth, tendon VIV effects were presented. Finally, conclusions were draw. 2. TLP Coupled Dynamic Analysis A six-DOF coupled dynamic equation of motion is expressed as follows
[M ] U + [C] U + [K ]{U} = {F}+ {Fm }

and added mass of slender members, [C] = damping matrix (6x6), potential damping + viscous damping + wave drift damping on hull + damping of slender members, [K] = stiffness matrix (6x6), hull hydrostatic stiffness (heave and roll/pitch) + stiffness due to tendons and risers, {F} = load vector (6x1) of first- & second-order wave loads + viscous loads + wind loads, and/or other applied loads, {Fm } = load vector (6x1) of tendon and riser tensions at the connected locations, {U } = unknown motion vector (6x1) in the sequence of surge, sway, heave, roll, pitch and yaw, respectively.

6-DOF Motion S olver

Line Dynamic S olver (T endon and Riser)

Motion Etc.
T ensions, Reaction Loads Etc.

Fig. 2 Coupling process chart

TLP hull/tendon/riser coupling process chart is illustrated in Fig. 2. The TLP force model is introduced as a nodal load component in a finite element (FE) model of tendons and risers. Among TLP hull and tendons or risers, forces are exchanged back and forth. It should be noted that this approach yields dynamic equilibrium between the forces acting on the TLP and tendons and/or risers at every time instant. 3. Tendon Mathieus Instability Zhang et al. (2002) investigated TLP tendon Mathieus instability under parametric excitation. After substituting boundary conditions and including hydrodynamic damping, a beam equation (tendon model) had been recast into a general Mathieus equation form as follows:
d2 f df +c + (a + b cos z ) f = 0 2 dz dz

(2)

{}

{}

(1)

where [M] = mass and inertia matrix (6x6), hull structural mass and inertia + hull added mass and inertia + mass

where (n l )[ EI (n l ) 3 + T0 (n l ) + w] , n is the tendon a= ml 2 mode, l denotes the tendon length, ml stands for the effective tendon mass per unit length, T0 represents tendon effective tension, and w means tendon under water weight per unit length. ml = Dh s + 0.25D 2 w , h is tendon wall thickness, s the density of steel.
387

 w the density of water, D the diameter of the tendon,

b= T ( n l ) 2 ml 2
Table 2 Maximum allowable tension variation

, T is tension variation amplitude,

0% 1% 5%

c= (4Cd w Du max ) (3m l ) , u max is maximum velocity of tendon transverse motion, Cd is hydrodynamic drag coefficient, z=t. The general stability diagram with damping effects as shown in Zhang et al. (2002) is adopted and reproduced as Fig.3.
Max Allowable Tension Variation (KN)
2.8

First Unstable Zone 0 2067 KN 6000 KN

Second Unstable Zone 495 KN 3617 KN 11400 KN

12000 First Unstable Zone 10000 Second Unstable Zone

8000

b (parameter in Mathieu's equation)

2.4

6000

2.0

4000

1.6

2000

1.2
0% o f Critical Damping 0% o f Critical Damping 1 o f Critical Damping % 1 o f Critical Damping % 5% o f Critical Damping 5% o f Critical Damping 1 o f C ritical Damping 0% 1 o f C ritical Damping 0%

0.8

0 0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

0.4

Damping %

0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

a (parameter in Mathieu's equation)

Fig. 4 Maximum allowable tension variation

Fig. 3 Stability diagram with and without damping

To demonstrate the importance of the damping on suppressing Mathieus instability, a case study has been followed and the TLP tendon parameter is given in Table 1.
Table 1 Tendon parameters
Pretension (mt) Tendon Diameter (O.D.) (m) Tendon Wall Thickness (mm) Tendon Length (m) Tendon Wet Weight (mt) 885.6 0.6604 26.29 826.5 58.2

Maximum allowable tendon tension variations around the first and second unstable zone with incident wave peak periods of 7.25 and 14.5 second are summarized in Table 2 and plotted in Fig. 4. a 0.25 and 1.0 are around the first and second unstable zones in Fig. 3. It has been found that the Mathieu instability will vanish fast with the presence of small damping in a higher-order unstable zone. Thus no tendon mode n>1 is considered in this study. From the Table 2, it has been noticed that maximum allowable tendon tension variations are very sensitive to the damping of tendon transverse motions. If damping is 0, the maximum allowable tendon tension variations for first and second unstable zone are 0 KN and 495 KN respectively for first and second unstable zones; if damping is 1% critical damping, maximum allowable tension variations dramatically
388

increase to 2067 KN and 3617 KN respectively for first and second unstable zones. In reality, the damping of tendon transverse motions is much higher than 1% of critical damping. Therefore tendon tension variations should be well within the allowable ranges. From Figs. 3 and 4, the higher-order regions of instability are more sensitive by damping than low-order ones. In Figs. 3 and 4, the higher-order regions of instability are fading away faster than those of the low-order ones by the same amount of damping. Since the low-order resonance zones are very important, the limited damping effects on low-order resonance zones may not be adequate to suppress Mathieus instability. Thus, careful examinations and assessments on the low-order resonance zones are recommended. Stability chart including damping effects has been generated by assuming regular incident waves. It still remains extremely challenging for developing a Mathieu stability diagram with damping effects for the irregular waves with the specified spectrum. 4. Tendon Springing and Ringing Tendon springing and ringing phenomena are due to the resonant pitching (rolling) and heaving motions at high frequencies. Fig. 5 presents one segment of the filtered high-frequency tendon tension time series. It clearly presents the characteristics of the springing and ringing, in which the springing is the highfrequency resonant tension response with the moderate amplitude while ringing is the transient and energetic event just like the Bell being knocked and

vibrating strongly.
2.E+07 Ringing 1.E+07 0.E+00

paper and complexity of the problem, only brief descriptions are highlighted as follows:
Springing
1.E+00 Gaussian Distribution 1.E-01 High Freq. Tension

Tension (N)

-1.E+07 -2.E+07 1140

Log10(Probability Density)

1160

1180

1200

1220

1240

1260

1280

1.E-02

Time (s)

1.E-03

Fig. 5 Tendon springing and ringing

No standard criteria has been officially agreed to differentiate the springing and ringing yet. However, parameters, such as extreme factor (extreme highfrequency tension / standard deviation high- frequency tension) and Kurtosis are frequently utilized to describe the springing and ringing events. In general, the springing is called if the extreme high-frequency tension does not exceed five to six times of the standard deviation and Kurtosis of the high-frequency tension is less or equal to 5.0 (Davis et al., 1994) while ringing is named if the extreme high-frequency tension exceeds seven times of the standard deviation and Kurtosis of the high-frequency tension is greater than 5.0 (Jefferys and Rainey, 1994). Figs. 6 and 7 illustrate typical probability distributions of springing and ringing events, and clearly indicate that both springing and ringing events are non-Gaussian distributions. It has been noticed that maximum extreme factor in Fig. 6 is slightly less than 6 while the extreme factor in Fig. 7 is close to 10.0.
1.E+00

1.E-04 0 2 4 6 8 10 Extreme Tension/STDEV

Fig. 7 Probability distribution of ringing tension

The ringing (the North Sea) has considerable impacts on extreme tendon strength. It is important to adequately account for the effects in the design either by physical model tests or hybrid method (minimum model tests + supplement numerical simulations model the model). The springing (the North Sea and Gulf of Mexico) has significant impacts on tendon fatigue life. Non-Gaussian effects should be treated properly. The springing is due to weak asymmetric waves while the ringing is due to strong asymmetric front-concave waves as shown in Fig. 8.
70 60

Wave Elevation (ft)

50 40 30 20 10 0 -10

Measured

Log10(Probability Density)

Gaussian Distribution High Freq Tension 1.E-01

1.E-02

-20 636

638

640

642

644

646

648

650

Time (s)
1.E-03 0 1 2 3 4 5 6

Extreme Tension/STDEV

Fig.8 Measured strong asymmetric wave profile

Fig. 6 Probability distribution of springing tension

The importance of investigating tendon springing and ringing responses are lieu in two folds, one is to quantify the impacts on extreme tendon strength and fatigue life; the other is to find what kinds of waves generate the springing and ringing and to develop reliable method to simulate. Significant progresses have been made during past 10 years (Davis et al., 1994; Jefferys and Rainey, 1994; Zou, 1997; Zou1998; Zou et al., 1999). Due to the limited length of the

To simulate strong asymmetric front-concave waves in numerical wave tank is extremely challenging in the near future. Therefore, it is important to generate this type of wave train in the wave basin for model tests. Zou et al. (2000) had developed the methodology and successfully generated strong asymmetric waves in random sea way by employing both time and crest distortions.

389

5. Tendon Tension Slacking and Wave overtopping Wave overtopping of the TLP column(s) and/or deck is becoming a serious design issue, especially after Hurricane Katrina and Rita and a mini-TLP being capsized. Prior to Hurricane Katrina and Rita, wave overtopping had brought our attentions when a three-column TLP physical wave basin model tests were performed in 1998 and tendon bottom tension slacking were observed in some runs. Fig. 9 illustrated the comparisons of the measured and simulated results by the model the model method (Zou et al., 2004). The wave overtopping on top of column was identified as the key factor to cause tendon tension slacking. Wave overtopping has been investigated in coastal engineering for fixed structures for many years. Up to date, it still remains extremely challenging to develop robust model for accounting for layer thickness, overtopping velocities, cumulative overtopping volume and mean overtopping rate. For floating platforms, it adds complexity and difficulty for proper simulation compared to fixed structures. Fig. 9 shows an example of intending to reproduce the measured results. Similarly, only a few key elements involved are briefly introduced as follows: Well proven and validated coupled dynamic time domain analysis program as base line. Additional to the above, the program should have the capability to model the model and account for wave overtopping: o Ability to de-composite the strong asymmetric wave as shown in Fig.9. Strong asymmetric is referred to both horizontal and vertical axis. o Capability to reconstruct and reproduce the measured strong asymmetric waves by the limited frequencies. o Validated weakly impact load model including weakly impact velocity and acceleration in both horizontal and vertical plans due to wave overtopping. 6. Tendon VIV Millennium eddy currents hit the Gulf of Mexico during the spring and summer of 2001. Since Millennium profile exceeded the existing design current velocities at the depth, a few platforms experienced strong tendon VIV and its impacts on operating. Strong excessive VIV on tendon will reduce its fatigue life, impact human comfortable level, and induce ex- cessive operating downtime. Thus, broad attentions have been draw and significant efforts have been made to suppress VIV impacts by devices and to develop reliable theory and tool to predict it.

4500 4000 Simulated Measured Measured Wave 3500 3000 2500 2000 1500 1000 500 0 -500 -1000 -1500

440 400 360 280 240 200 160 120 80 40 0 -40 520 540 560 580 600 620 640 660 680 700 720 740

Tnedon Tension (kips)

Time (s)

Fig. 9 Simulated and measured tendon tension, and strong asymmetric wave

As illustrated in Fig. 10, towing tests were carried out at the well-known Offshore Technology Research Center (OTRC) in 2004, with TLP hull, all tendons and top tensioned risers (TTRs) attached without any truncation in length. This was the first time in the world to tow the entire platform in the wave basin. The advantages of towing the entire model are as follows: It is the closest model to the prototype. Account for the hull/tendon coupling effects. Mean tension effects on different tendons due to mean offsets are inherently built in. The damping effects from hull and TTRs are also inherently included.

Fig. 10 Underwater View of Towing Tests Setup

The prototype field is in 4,000 ft water depth. Towing tests consist of towing model in still water and in waves. The towing tests are intended to simulate loop current conditions in a conservative and controllable way (uniform current from top to bottom, speed up to 6.0 ft/s). The measured natural periods of heave and roll/pitch are 3.75 s and 3.35 s / 3.38 s respectively. The measured bending stiffness of tendon model pipe and axial spring were employed in the prediction. Tables 3 to 5 document the comparisons of natural periods of vibrations, Re and St numbers of the prototype, ideal model and actual model. It can be seen that both Re and St numbers are very close between the ideal and actual models. However, significantly differences are found w.r.t.

390

Wave Elevation (ft)

320

those of the prototype. For low and high mean tendon tensions in 30 degree towing, possible modes being excited by tendon VIV are shown in Tables 6 and 7, and the predicted and measured tension RMS are compared in Tables 8 and 9. In general, the predicted tension RMS are significantly higher that those measured, which might be caused by the difference between the predicted mean tensions and the measured. It is also notice that good agreements have been reached at speed 3 ft/s for both low and high mean tensions and at speed 5 ft/s for high mean tension. The measured data are filtered out low-frequency components (period longer than 28.2 second) since they are not due to tendon VIV based on Table 1 Actual Model.
Table 3 Natural periods of vibrations
M ode Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Prototype (sec) 29.4 14.7 9.8 7.3 5.8 4.8 4.1 3.5 3.1 2.8 2.5 2.3 2.1 1.9 Ideal M odel (sec) 29.4 14.7 9.8 7.3 5.8 4.8 4.1 3.5 3.1 2.8 2.5 2.3 2.1 1.9 Actual M odel (sec) 28.2 12.7 7.4 4.8 3.3 2.4 1.8 1.4 1.2 1.0 0.8 0.7 0.6 0.5

Current (ft/s) 1 2 3 4 5 6

M ean Top Tension (kips) 1870 1946 2161 2532 3091 3740

M ode

Period (s) 13.36 7.66 7.43 4.68 4.44 3.03

2 3 3 4 4 5

Table 8 Comparisons of predicted and measured RMS tensions of low mean tension tendon
C rren u t (ft/s) 1 2 3 4 5 6 M e od P d R SA erio M /D P icted red M easured (s) R STensio R STen n M n M sio (kip s) (kip s) 13.42 0.89 18.2 6.5 7.76 0.86 37.9 17.8 7.64 0.92 43.5 38.6 4.85 0.96 83.7 43.9 3.31 0.96 131.4 88.9 3.22 1 140.6 122.3

2 3 3 4 5 5

Table 9 Comparisons of predicted and measured RMS tensions of high mean tension tendon
Current (ft/s) 1 2 3 4 5 6 Mode Period (s) 13.36 7.66 7.43 4.68 4.44 3.03 RMS A/D Predicted Measured RMS Tension RMS Tension (kips) (kips) 17.9 5.7 47.6 18.3 38.6 42.2 87.3 43.5 74.2 86.3 132.9 114.6

Table 4 Re number comparisons

Current (ft/s) 1 2 3 4 5 6 Prototype 2.02E+05 4.05E+05 6.07E+05 8.10E+05 1.01E+06 1.21E+06 Ideal Model Actual Model 3.05E+02 3.41E+02 6.11E+02 6.83E+02 9.16E+02 1.02E+03 1.22E+03 1.36E+03 1.53E+03 1.70E+03 1.83E+03 2.04E+03

2 3 3 4 4 5

0.88 0.96 0.87 0.98 0.9 0.97

Table 5 St number comparisons

Current (ft/s) Prototype 1 0.24 2 0.24 3 0.24 4 0.24 5 0.24 6 0.24 Ideal Model Actual Model 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17 0.17

Table 6 Modes and periods of low mean tension tendon

Current (ft/s) 1 2 3 4 5 6

Mean Top Tension (kips) 1849 1861 1967 2196 2515 2889

Mode

Period (s) 13.42 7.76 7.64 4.85 3.31 3.22

2 3 3 4 5 5

Table 7 Modes and periods of high mean tension tendon

In order to investigate the contributions from different tendon modes, the high pass filtering progresses to isolate each mode contribution have been undertaken. Figs. 11(a) to 11g illustrate the high pass filtered time series of the measured tendon tension up to mode #7 per Table 3 Actual Model. In this test, the towing speed is 5.0 ft/s (no wave and wind). Fig. 11(a) denotes the first mode high pass filtered time series excluding mean and low-frequency components (cut-off period 28.2 s). However, it did include all possible modes due to tendon VIV. Fig. 11(b) stands for second mode high pass filtered time series, which excludes first mode contribution by setting cut-off period 12.7 s. High pass filtering processes were repeating up to tendon mode #7 as shown in Fig. 11(g). From Figs. 11(a) to 11(g), a few observations are summarized as follows: It is clear that multi-mode vibrations contribute to the total tendon tension dynamics. In this case, mode #2, #4 and #6 contribute more, especially mode #2. Mode numbers higher than 6 have virtually no contribution.

391

200

100 0 -100 -200 50 70 90 110 130 150

Time (s)

Fig. 11(a)
200

First mode high pass tension time series

excessive operating downtime and human comfort. The maxi- mum towing speed 6.0 ft/s represents uniform current profile, which is believed to be conservative compared to typical GoM eddy current. Further investigations are underway to explore why and thus might result better platform design to suppress tendon VIV impacts. 7. Conclusions This paper introduces hydrodynamic aspects of in-place TLP tendon design and analysis with emphases on TLP hull/tendon/riser coupled dynamic analysis; tendon Mathieu instability; tendon springing and ringing responses; tendon bottom tension slacking due to wave overtopping; and tendon vortex induced vibration (VIV). The progress in these aspects has been highlighted and discussed. Future works are also recommended. References

Tension (kips)

Tension (kips)

100 0 -100 -200 50 70 90 110 130 150

Time (s)

Fig. 11(b) Second mode high pass tension time series

200

Tension (kips)

100 0 -100 -200 50 70 90 110 130 150

Time (s)

Fig. 11(c) Third mode high pass tension time series

200

[1]

[2]

100 0 -100 -200 50 70 90 110 130 150

[3]
Time (s)

Fig. 11(d) Fourth mode high pass tension time series

200

[4]

100 0 -100 -200 50 70 90 110 130 150

[5]

[6]

Time (s)

Fig. 11(e) Fifth mode high pass tension time series

200

[7]

100 0 -100 -200 50 70 90 110 130 150

[8]

Time (s)

[9]

Fig. 11(f) Sixth mode high pass tension time series

200

Tension (kips)

[10]

100 0 -100 -200 50 70 90 110 130 150

Time (s)

[11]

Fig.11(g) Seventh mode high pass tension time Ssries [12]

The platform behaves well. No platform response was observed that would lead any concern for
392

Davies, KB, Leverett, SJ and Spillane, MW (1994). Ringing Response of TLP and GBS Platforms, Proc 7th Int Conf Behavior Offshore Structures, Vol 2, pp 569-585. Jefferys, ER and Rainey, RCT (1994). Slender Body Models of TLP and GBS Ringing, Proc Int Conf Behavior Offshore Structures, Vol 2, pp587-605. Kim, CH, Kim, MH, Liu, YH and Zhao, CT (1994). Time Domain Simulation of Nonlinear Response of a Coupled TLP System in Random Seas, Int J Offshore and Polar Eng, ISOPE, Vol 4, No 4, pp 284-291. Leverette, S, Rijken, O, Dooley, W and Thompson, H (2003) Analysis of TLP VIV Responses to Eddy Currents, Proc OTC 15289, Houston, TX. Ma, W, Lee, MY, Zou, J and Huang, EW (2000) Deepwater Nonlinear Coupled Analysis Tool, Proc OTC 12085, Houston, TX. Natvig, B.J. (1994). A Proposed Ringing Analysis Model for Higher Order Tether Response, Proc 4th Int Offshore and Polar Eng Conf, Osaka, ISOPE, Vol 1, pp 40-51. Ormberg, H and Larsen, K (1998) Coupled Analysis of Floater Motion and Mooring Dynamics for a Turret-Moored Ship, Appl Ocean Res, Vol 20, pp 55-67. Paulling, JR and Webster, WC (1986). A Consistent, Large-Amplitude Analysis of the Coupled Response of A TLP and Tendon System, OMAE 1986. Zhang, LB, Zou, J, and Huang, EW (2002). Mathieu Instability Evaluation for DDCV/Spar and TLP Tendon Design, Proc of the 11th Offshore Symposium, SNAME, Houston, TX. Zou, J (1997). Investigation of Slowly-Varying Drift Motion and Springing and Ringing of Tension Leg Platform System in Nonlinear Irregular Waves, Ph D Dissertation in Ocean Eng, Texas A & M University, December, 1997. Zou, J (2003), TLP Hull/Tendon/Riser Coupled Dynamic Analysis in Deepwater, pp. 160-166, ISOPE 2003 Conf., Honolulu, Hawaii, USA, June 2003. Zou, J and Kim, CH (1996). Experimental Study of Impacting Wave Force on Vertical Truncated Cylinder, Int J Offshore and Polar Eng, ISOPE, Vol 6, No 4, pp 291-293.

Tension (kips)

Tension (kips)

Tension (kips)

[13] Zou, J. and Kim, C.H. (2000). Generation of Strongly Asymmetric Wave in Random Seaway, Proc. 11th Int. Conf. of Offshore and Polar Eng. Vol. 3, pp. 95-102. [14] Zou, J, Huang, E.W., and Kim, C.H. (1999). Nonlinear and Non-Gaussian Effects on TLP Tether Responses, Proc 9th Int. Offshore and Polar Eng Conf, Brest, France, Vol. 1, pp. 315-324.

[15] Zou, J, Ormberg, H and Stansberg, CT (2004) Predictions of TLP Responses, Model Tests Vs. Analysis, Proc OTC 16584, Houston, TX. [16] Zou, J, Xu, Y and Kim, CH (1998). Ringing of ISSC TLP due to Laboratory Storm Seas, Int J Offshore and Polar Eng, Vol 8, No 2, pp p81-89.

393