Proceedings of the 26th International Conference on Offshore Mechanics and Arctic Engineering OMAE2007 June 10-15, 2007, San Diego, California, USA

Transcription

1 Proceedings of the 26th International Conference on Offshore Mechanics and Arctic Engineering OMAE2007 June 10-15, 2007, San Diego, California, USA OMAE NUMERICAL SIMULATION OF THE DEPLOYMENT OF A HYBRID ROV OPTICAL FIBER TETHER Mark A. Grosenbaugh Woods Hole Oceanographic Institution Woods Hole, MA Brian Bingham Olin College Needham, MA Louis L. Whitcomb The Johns Hopkins University Baltimore, MD Jason I. Gobat APL University of Washington Seattle, WA Chris Young Space and Naval Warfare Systems Center San Diego, CA ABSTRACT This paper describes a numerical code for simulating the dynamics of an unmanned, underwater vehicle system that is self-propelled and tethered to a surface ship through an optical fiber tether. The vehicle, called a hybrid ROV, uses a buffered, single-mode optical fiber with a maximum working load of 1.7 N for communication and data transmission. The vehicle is designed to go to the deepest parts of the ocean and for exploring beneath the Arctic ice cap. The optical fiber tether is stored in a pair of canisters, one mounted on the vehicle and one mounted on a garage that is lowered from the ship. The canisters each hold 20 km of fiber, which is pulled out during operations when the tension at the canister reaches a threshold value, which is set to the maximum working load of the fiber. The numerical simulation is based on the twodimensional version of WHOI Cable, a finite-difference solver of the cable equations that includes bending stiffness to model low-tension effects. A velocity/tension based payout algorithm was incorporated into the code to model the behavior of the canisters. In the payout model, the payout velocity is set equal to zero below the threshold tension and varies linearly with tension above the threshold value up to a maximum pay out velocity. Hydrodynamic drag models for axial and normal fluid loading, whose values are a function of Reynolds number, are used to calculate local drag coefficients of the optical fiber. Examples of the vehicle being lowered to the sea bottom in uniform and shear currents are used to demonstrate the capabilities of the code and the performance of the tether and payout system. INTRODUCTION This paper describes and demonstrates a numerical program that was developed for studying the dynamics of a new type of underwater vehicle. The new vehicle, which is being developed for ocean research, combines battery power systems used by autonomous underwater vehicles (AUV) and fiber optic tether technology used by remotely operated vehicles (ROV). The result is a hybrid remotely operated vehicle (HROV) that is essentially free-swimming except for a high bandwidth data link back to the ship (Bowen et al., 2004). Because the vehicle is self-powered, the tether can be a single optical fiber with a diameter that is nearly two orders of magnitude smaller than a typical ROV tether. In addition, the vehicle can be disconnected completely from the tether and operated as an AUV (Figure 1). The HROV will provide access to areas of the ocean that, up until now, are virtually unexplored. These include the deep

2 Figure 1. Engineers at the Woods Hole Oceanographic Institution are developing a Hybrid Remotely Operated Vehicle (HROV) with an optical fiber tether that will have a depth range of 11,000 m. (Illustration by Jack Cook, WHOI) trenches, which are some of the most active earthquake zones on earth, deep transform faults along mid-ocean ridges, and under the polar ice caps. The deepest point of the ocean, the Mariana Trench with a maximum depth of 10,911 m, has been visited only twice in history by the manned bathyscaphe Trieste in 1960 (Picard and Dietz, 1962) and the Japanese remotely operated vehicle KAIKO in 1995 (Mikagawa et al., 1999). The Trieste made its last dive in 1983, and KAIKO was lost at sea in Currently, there are no vehicles that are capable of reaching the deepest part of the ocean. HROV, which will fill this void, has the added benefit of being compact enough for quick deployment by a small crew from virtually any ship in the world. The unique feature of the HROV is the optical-fiber tether and the method used to pay it out during operations. The main tether is a commercial-grade coated optical fiber with an outside diameter of 245 µm (OFS TrueWave RS optical fiber). The optical fiber tether is spooled onto two canisters for vehicle deployment (Figure 2). One of the canisters is attached to a depressor weight that is lowered with the HROV from a ship by an 8-mm diameter steel armor cable. The other canister is attached to the HROV itself. The canisters can each hold up to 20 km of fiber, which is pulled out during vehicle Figure 2. Deployment of the HROV (1) vehicle is lowered below strong surface currents. (2) vehicle descends with optical fiber paying out from canisters fixed to the depressor and the vehicle frame. (3) vehicle continues to work on sea bottom, paying out fiber as needed. (4) vehicle drops fiber end and returns to ship. (5) optical fiber is recovered on board surface ship. (Illustration by Don Peters, WHOI) operations when the tension at the canister reaches a threshold value. The threshold is set equal to a value that is below the maximum working load of the fiber. Numerical simulations of the HROV system involve computing low-tension cable dynamics combined with a tension based pay-out system that can increase the deployed tether length from 0 m at the start of the simulation to potentially 40 km at the end of the simulation. In addition, the simulation must account accurately for the low Reynolds number drag of the optical fiber. Previous simulations of ROV tethers involved large-diameter, fixed-length cables (Grosenbaugh et al., 1993; Buckham and Nahon, 2001). These simulations take into account low-tension effects through the inclusion of bending stiffness in the equations of motion. Cable pay out from a winch at a prescribed pay-out velocity has been simulated quasi-statically through boundary conditions (Burgess, 1991; Sun et al., 1994) and dynamically through the equations of motion (Burgess, 1993). While we follow the dynamic formulation for pay-out velocity, we are also able to incorporate a functional pay-out velocity into the

3 Figure 3: Fiber payout velocity as a function of fiber tension. The shape of the payout function used in the simulation is user defined. For the present simulations, the threshold tension is 0.1 N, the slope of the payout function above the threshold tension is 1 N/m/s, and the maximum payout velocity is 5 m/s. equations of motion that varies both with time and instantaneous tension. For the optical fiber canisters used with the HROV, this function specifies a threshold tension, below which pay-out velocity is zero. We have also modified the hydrodynamic drag functions to take into account low Reynolds number effects. A typical Reynolds number for an optical fiber tether operating in the ocean is less than 100. In the next section, we describe our numerical simulation for the dynamics of low-tension cables. We include in this discussion a complete description of the tension-based payout model for the optical fiber canisters. Section 3 contains a description of the HROV vehicle and the material characteristics of the optical fiber tether. Section 4 contains a discussion of previous laboratory measurements of longitudinal and normal drag forces on cylinders at very low Reynolds numbers. Here, we present the functional forms of the tangential and normal drag coefficients used in the simulation. Simulation results that demonstrate the capabilities of our numerical model are also given in Section 4. This is followed by discussion of the simulation results in Section 5. CABLE MODEL The core of the numerical simulation is built around WHOI Cable, which is an all-purpose software package for analyzing the statics and dynamics of oceanographic cable systems (Gobat and Grosenbaugh, 2006). The software can be used for analyzing surface and subsurface single-point mooring problems, multi-leg and branched array systems, and towing and drifting problems. The numerical model consists of a set of finite difference equations that include the effects of geometric and material nonlinearities, bending stiffness for seamless modeling of slack cables (including snap loading), and a model for the interaction of cable segments with the sea bottom. The main feature of the program is a novel time integration algorithm that produces accurate, stable, and robust solutions to dynamic problems. Adaptive time stepping, adaptive gridding, and adaptive relaxation greatly aid users in obtaining solutions quickly. User specified forcing can include waves, currents, wind, ship speed, and vehicle thrust. In the real system, optical fiber uncoils from canisters fixed at the two ends of the optical fiber tether. For the simulation, the payout of the optical fiber is incorporated in the equations of motion through the material derivative (Burgess, 1993). A payout function that depends on the instantaneous tension at the ends of the optical fiber tether is specified and the simulation is allowed to proceed. The general parameters of the payout function are the maximum payout velocity, the slope of the transition between zero payout velocity and maximum payout velocity, and the threshold payout tension (Figure 3). For our application we have set the maximum payout velocity equal to 5 m/s (compare this to the nominal HROV descent velocity of 0.5 m/s). The slope of the payout function is 1 m/s/n. The threshold tension is set equal to 0.1 N. SYSTEM PARAMETERS HROV is called a hybrid vehicle because it is selfpowered and can operate independently as an AUV yet it can be tethered to a ship for real-time communication and data transmission like an ROV. Because the tether does not carry any electrical power, its diameter can be very small in our case about 250 µm. The vehicle itself weighs about 2100 kg in air and is nearly neutrally buoyant in water. It has two aft thrusters, two vertical thrusters, and one lateral thruster. It has a maximum speed of 1.5 m/s. Power is delivered from dual 6 kw rechargeable lithium-ion battery packs similar to those used in the ABE vehicle (Bradley et al., 2000) but with increased capacity. The total payload capacity is 25 kg. The vehicle can be transformed on a ship s deck from AUV-mode used for wide-area ocean surveys to ROV-mode for close-up investigation and sampling of seafloor rocks and organisms. Technologies that will be employed to allow for the increased operation depth include using high strength syntactic foam and ceramic pressure housings for electronics. In the present simulations, the effect of the vehicle enters only through the user-inputted descent velocity. In our simulations, the vehicle descends at a constant velocity of 0.5 m/s to the sea bottom where it then holds position for the rest of the simulation. The HROV tether is nearly as thin as a strand of human hair and not much stronger (Table 1). Fully deployed it will stretch 40 km, but because of its small diameter it will have very low drag. The biggest issue for the HROV is whether the tension in the deployed optical-fiber tether exceeds the working strength corresponding to long (> 4 hours) or short (< 4 hours) term events. Exceeding the working strength of the optical fiber can greatly increase the probability of failure (Castilone et al., 2000). Another issue is whether the optical fiber interacts with the sea bottom during the deployment. If this occurs, our assumption is that the tether will break. The

4 Material Property TrueWave RS Optical Fiber Diameter (m) Mass per length (kg/m) Depth (m) Current Speed (m/s) Wet weight per length (N/m) (sea water) Safe working strength (N) (long-term event > 4 hours) Safe working strength (N) (short-term event < 4 hours) Proof test strength N 2.82 N 8.47 N Table 4: Current Profile. purpose of the numerical simulation during design will be to answer these questions for a given mission. Table 1: Specifications for TrueWave RS optical fiber. Material Property Steel-Armor Cable Diameter (m) Mass per length (kg/m) 1.12 Wet weight per length (N/m) (sea water) 8.89 Normal drag coefficient 1.5 RESULTS The accuracy of the simulation depends on the hydrodynamic drag model of the optical fiber. In practice, the maximum current acting on the optical-fiber tether can be reduced by first lowering the depressor weight, fiber canisters, and HROV as a single package on an 8-mm diameter steel armor cable to a depth that is below strong currents (say < 40 cm/s). This depth is typically below the ocean mixed layer, which can extend down from the surface several hundred meters. If we restrict the current values to between 0 and 40 cm/s, then the Reynolds numbers are limited to values between 0 and 100. Pao et al. (2000) made measurements of the axial hydrodynamic force on yawed cylinders at very low Reynolds numbers. They found that the data followed an inverse power law such that: Axial friction coefficient C f = ARe (1) 1 t Table 2: Specifications for standard 0.32-inch oceanographic steel armor cable. Material Property Spherical Depressor Weight Diameter (m) 1.0 Mass (kg) 500 Wet weight (N/m) (sea water) 4370 N Drag coefficient 1.5 Table 3: Specifications for spherical depressor weight. where C f is the axial friction coefficient and Re t is the Reynolds number based on the tangential velocity. The proportionality constant that best fits all the data for Reynolds number less than 100 is A = 9.1, which is the same value published by Karnoski (1991) after he analyzed Pao et al. s data, which was unpublished at that time. Because of the possibility of vortex shedding and strumming of the optical fiber in currents, the normal drag coefficient is more difficult to define in functional form. However, recent numerical computations of flow around a stationary cylinder have produced normal drag coefficients for Reynolds numbers ranging from 0 to 100 that closely match previous experiments (Sheard et al., 2005). These data can be fit with the following form: C = C Re + C (2) 1 d 1 n 2 where C d is the normal drag coefficient and Re n is the Reynolds number based on the normal velocity. The proportionality constants C 1 and C 2 determine the value of the linear and quadratic drag terms respectively. For the data in

5 Sheard et al. (2005), the best fit is C 1 = 17 and C 2 = 1.2. Strumming effects, which would occur above Reynolds number 40, could increase C d though the effects would be mitigated by the tether s long length, low tension, and typically curved geometry, which act to prevent lock-in at a natural mode. The remaining input parameters for the simulations are the material properties of the steel armor lowering cable and the depressor weight that hangs just below the ship and houses the upper optical fiber canister. These are provided in Table 2 and Table 3. The depressor weight is modeled as a point sphere for ease of computations. The current profile for the simulations is given in Table 4. The current decreases from 0.50 m/s at the surface to 0.25 m/s at a depth of 250 m. The current velocity is then uniform from 250 m to the sea bottom at 1000 m and has a speed of 0.25 m/s. The first simulation (CASE 1) begins at t = 0 with the depressor weight and HROV hanging from the steel armor lowering cable at a depth of 250 m in static equilibrium. This depth was chosen to be below the strong shear currents in the surface layer. The HROV descends to the sea bottom (at 1000 m) at a constant rate of 0.5 m/s (Figure 4). This process takes 1500 s. After touchdown, the optical fiber tether continues to pay out due to the current drag. Eventually at t = 3600 s, enough optical fiber has payed out so that so that a portion of it lies on the sea bottom (Figure 5). Figure 4. CASE 1: Numerical simulation of the deployment of the HROV with the vehicle and pay-out canisters lowered to a depth of 250 m to avoid the strong shear currents in the surface layer. The current velocity below 250 m is uniform with a speed of 0.25 m/s. Figure 6. CASE 2: Numerical simulation of the deployment of the HROV with the vehicle and pay-out canisters lowered to a depth of 100 m. The current speed at a depth of 100 m is 0.4 m/s and then decreases to 0.25 m/s at a depth of 250 m. From 250 m to 1000 m, the current is uniform velocity. Figure 5. CASE 1: Close-up of sea bottom touchdown during simulation with uniform current conditions of 0.25 m/s. Vehicle touchdown occurred at t = 1500 s. After t = 3600 s, the optical fiber is laying on about 50 m of the sea bottom. Figure 7. Close-up of sea bottom during CASE 2 simulation with sheared current in the upper surface layer. Vehicle touchdown occurred at t = 1800 s. After t = 3600 s, the optical fiber is just above the sea bottom.

6 The second simulation (CASE 2) begins at t = 0 with the depressor weight and HROV hanging from the steel armor lowering cable at a depth of 100 m in static equilibrium. This places the top canister in the region of strong shear - the current speed at 100 m is 0.40 m/s. The HROV then descends to the sea bottom at a constant rate of 0.5 m/s (Figure 6). This process takes 1800 s. After touchdown, the optical fiber tether continues to pay out due to the current drag. After 3600 s, all of the optical fiber is still suspended above the sea bottom (Figure 7). The length of deployed optical fiber tether as a function of time for each of the simulations is shown in Figure 8. The initial slope of the curves is dominated by the descent speed of the HROV. After the vehicle touches down on the sea bottom (at t = 1500 s for CASE 1 and t = 1800 s for CASE 2), the slope of the curves decreases slightly though optical fiber continues pay out due to current drag. Figure 9 shows the distribution of tension along the length of the optical fiber tether at different times for CASE 1 (top panel) and CASE 2 (bottom panel). The curves are separated in time by intervals of 600 s. The tether coordinate is referenced to the extreme end of the optical fiber inside the canister of the HROV. For these simulations, the initial length of the optical fiber is 6000 m with the length being equally divided between the two canisters. The discontinuities in the slopes of the curves occur where the optical tether exits the canisters and, thus, bound the portion of the fiber that has been deployed. The leftmost discontinuity is at the HROV and the rightmost discontinuity is at the depressor weight just below the ship. The leftmost discontinuity moves to the left with time, and its location indicates how much optical fiber has been pulled out from the HROV canister (or conversely how much fiber remains in the HROV canister). Likewise, the location of the rightmost discontinuity (which moves to the right with time) indicates how much optical fiber has been pulled out from the top-side canister. The higher tensions in the optical fiber at the HROV for time t < 1500 s (CASE 1) and t < 1800 s (CASE 2) are due to the descent of the vehicle. Once the vehicle touches down on the sea bottom, the optical fiber tension at the HROV are reduced below the values measured at the topside canister. DISCUSSION Shear current in the upper part of the water column has the beneficial effect of lifting optical tether upward in the water column and delaying touchdown of the tether on the sea bottom. For CASE 1, a 50 m portion of the tether was lying on the sea bottom after 3600 s. For CASE 2, none of the tether was contacting the sea bottom after this time, though continuation of the simulation showed roughly 50 m of contact at t = 4000 s. Under the conditions used for the CASE 1 and CASE 2 simulations, it appears that equilibrium will never be obtained (except when all the optical fiber is deployed from both canisters, at which point the fiber would break in real life due to the sudden increase in tension). This is a conjecture based on the extrapolation of the data in Figure 8 and shown Figure 8. Length of deployed optical tether for CASE 1 and CASE 2. The discontinuity in the slope of each curve represents vehicle touchdown after which the tether continues to pay out due to current drag, but at a slower rate. Figure 9. Tension in the optical fiber tether as a function of tether coordinate. The curves in each plot, which correspond to different times, include fiber that is still contained in the canisters. The discontinuities in the slopes of the curves mark the ends of the tether and bound the portion of the fiber that has been deployed. The leftmost discontinuity is at the HROV and the rightmost discontinuity is at the depressor weight. The highest tensions occur at the HROV at times that precede the touchdown of the vehicle on the sea bottom.

7 by further simulations of up to 14,400 s. The allowable safe fatigue stress level for an optical fiber is 33% of the proof test strength for a time period of 4 hours and 20% of the proof test strength for longer application times (Castilone et al., 2000). The proof test strength of our fiber is 8.47 N, thus a 33% level is 2.82 N and a 20% level is 1.69 N (Table 1). For both CASE 1 and CASE 2, the tensions are always less than the allowable fatigue stress level for long periods of time. The maximum tension in the optical fiber tether occurs at the HROV and it occurs during the descent of the vehicle from the depressor weight to the sea bottom. This is an artifact of the payout velocity being higher at the vehicle than at the depressor. The maximum tension is always less than 0.6 N for both cases. It is possible to obtain optical fiber with double the proof test strength (i.e. ~17 N), which would greatly improve fatigue resistance of the optical fiber for the deepest descents. The results presented in this paper were meant to demonstrate the capabilities of the numerical method. Simulations and future full-scale trials in weaker currents and shear currents that extend to the sea bottom, which are more realistic in terms of what is encountered in the ocean, will hopefully show that an optical fiber tether can maintain a position above the sea bottom for the duration of a mission and operate with tensions that do not impact its fatigue life. ACKNOWLEDGEMENTS Support for this research was provided by the National Science Foundation (NSF) under grant number OCE Funding for the development of WHOI Cable software was provided by Office of Naval Research (ONR) under grants N J-1269 and N REFERENCES Bowen, A., Yoerger, D., Whitcomb, L. and Fornari, D., Exploring the Deepest Depths: A Novel Light-Tethered Hybrid ROV for Global Science in Extreme Environments, MTS Journal, Vol. 38, pp Bradley, A., Duester, A.R., Liberatore, S.P., and Yoerger, D.R., Extending the Endurance of an Operational Scientific AUV Using Lithium-Ion Batteries, Proc. of UUVS 2000, Southampton, UK. Buckham, B.J. and Nahon, M., Formulation and Validation of a Lumped Mass Model for Low-Tension ROV Tethers, Int. J. Offshore and Polar Eng., Vol. 11, pp Burgess, J.J., Modeling of Undersea Cable Installation with a Finite Difference Method, Proc. 1 st Int. Offshore and Polar Eng. Conf., Edinburgh, UK, Vol. 2, pp Burgess, J.J., Bending Stiffness in a Simulation of Undersea Cable Deployment, Int. J. Offshore and Polar Eng., Vol. 3, pp Castilone, R.J., Glaesemann, S.G., Hanson, and T.A., Extrinsic Strength Measurements and Associated Mechanical Reliability Modeling of Optical Fiber, Proc. 16 th Annual National Fiber Optic Engineers Conf., Denver. Gobat, J.I. and Grosenbaugh, M.A., Time-Domain Numerical Simulation of Ocean Cable Structures, Ocean Engineering, Vol. 33, pp Grosenbaugh, M.A., Howell, C.T., and Moxnes, S., Simulating the Dynamics of Underwater Vehicles with Low-Tension Tethers, Int. J. Offshore and Polar Eng., Vol. 3, pp Karnoski, S., A Model for the Tangential Hydrodynamic Drag on Small Diameter Cables Based on Laboratory Measurements, Naval Civil Eng. Lab. Tech. Memorandum, TM No. 43P Mikagawa, T., Fukui, T., and KAIKO Operation Team, ,000-Meter Class Deep Sea ROV KAIKO and Underwater Operations, Proc. 9 th Int. Offshore and Polar Eng. Conf., Brest, France, Vol. 2, pp Pao, H.P., Ling, S.C., and Kao, T.W., Measurement of Axial Hydrodynamic Force on a Yawed Cylinder in a Uniform Stream, Proc. 10 th Int. Offshore and Polar Eng. Conf., Seattle, Vol. 3 pp Picard, J. and Dietz, R.S Seven Miles Down: The Story of the Bathyscaph Trieste, Longmans, Green and Company, London, ISBN Sheard, G.J., Hourigan, K., and Thompson, M.C., Computations of the drag Coefficients for Low-Reynolds- Number Flow Past Rings, J. Fluid Mech., Vol. 526, pp Sun, Y., Leonard, J.W., and Chiou, R.B., Simulation of Unsteady Oceanic Cable Deployment by Direct Integration with Suppression, Ocean Engineering, Vol. 21, pp