Moving an ice mountain.
When an iceberg presents an immediate danger it is towed, but the methods available to accomplish this are both difficult and unreliable. The most common technique, and one that has been practiced for some time, is to pass a buoyant hawser around the iceberg to form a bight. Unfortunately, the bight tends to slip over the top of the iceberg and the iceberg tends to roll, limiting the possible towing load. Moreover, the technique is impractical in the case of large icebergs because of the complications of handling very long, multiple lengths of tow hawser. More complex systems using nets or multiple lines have also proved impractical under severe weather conditions.
An alternative to traditional towing methods employs a combined drill-and-anchor tool embedded in the iceberg by a remotely operated vehicle (ROV). Two tests of a prototype tool were successfully conducted on icebergs off the coast of Newfoundland. Long-term towing loads of over 25 tons were achieved in both cases. The results of these prototype tests have been used to design a full-scale tool.
Drilling is accomplished relatively quickly by ultrahigh-pressure water-jets mounted on a rotating swivel. When the anchor section of the tool has penetrated the ice, it is frozen in place by liquid carbon dioxide fed to the anchor and allowed to vaporize inside it. The tool is released from the ROV, which returns to the workboat, and a steel tow cable is then connected to a polypropylene hawser attached to the anchor section. The cable is let out until the workboat is 1500 meters from the iceberg, before towing begins.
Load requirements for towing are determined by the winds and currents whose effects on iceberg motion need to be overcome. The drag coefficient of 50,000- to 22-million-ton icebergs in the Labrador Sea has been estimated at 0.55. These icebergs are generally nontabular, with an average weight of 900,000 tons and a keel area of 36,000 square meters. Workboats in the area are typically capable of a 50-ton bollard pull for continuous towing. These values imply that the average iceberg can be towed at a speed of 0. 14 knots.
Iceberg drift velocities due to winds and currents in the Labrador Sea are typically in the range of 0.2 to 1.0 knots. Thus, a workboat should be able to significantly alter the course of the average iceberg; a 24-hour tow, for example, would move an iceberg 7 kilometers. The tow force varies linearly with keel area, which in turn varies with the weight of the iceberg. A 9-million-ton iceberg, which is 10 times the average size, could be towed at the same speed by five workboats. Workboats of the type available off the eastern shore of Canada are therefore capable of significantly deflecting almost any iceberg, provided a means of attachment is available.
Ready to Drill
The design of an anchoring tool is constrained by logistics and by the loads required for effective towing. To begin with, installing the anchor from an ROV severely restricts the thrust available for drilling a hole; in the presence of currents and wave action, most of the vehicle's power must be used to hold the drill stationary relative to the ice face. In addition, drilling and anchoring must be accomplished quickly to minimize the risk to the ROV from unstable icebergs.
The drill-and-anchor tool, 4 inches in diameter and 15 feet long, consists of a threaded steel tube containing a rotary waterjet drill head followed by a cryogenic anchor. The use of an ROV dictates that the drill used to prepare the anchor hole be rapid and simple to manipulate. An ultrahigh-pressure, rotary waterjet drill requires less than 100 pounds of thrust. This is critical, since the thrust available from existing ROVs is limited. The tool is powered by a pump on the workboat.
The jet nozzles and rotation motor can be mounted in a single compact unit at the front of the integrated drill-and-anchor tool. This design allows drilling and anchoring to be performed in one operation, thereby eliminating the difficulty of drilling the hole, removing the drill, and reentering the hole with an anchor. Since only the waterjets rotate, the ROV is simply required to hold the tool steady and provide a constant thrust of about 20 pounds.
The waterjets are produced by forcing water at 35,000 psi through four small (0.012- to 0.0 18-inch diameter) sapphire nozzles. The water velocity as it leaves the nozzles is about 2000 feet per second. In tests, waterjet cuts in ice show a smooth surface. The ice is apparently melted by the heat generated as the jet impacts and dissipates its energy. The loads in the impact region are two to three orders of magnitude greater than the compressive strength of the ice. Stagnation pressures here are large enough to cause pressure melting, while the plastic deformation of the ice nearby generates enough heat to melt it. The water flowing from a hole cut by a waterjet contains very little ice, indicating that the cutting mechanism itself is melting.
The tip of the drill is a nonrotating, beveled tube that controls the hole diameter. The drill is fed at a constant thrust so that it moves forward as the ice ahead of it is removed. A thrust of 20 to 100 pounds is sufficient to feed the drill; higher thrusts tend to reduce the drilling rate by jamming the rotary head. In tests using a 4-inch drill operating at 20,000 psi on commercial ice blocks, drilling rates of up to 18 inches per minute were achieved.
The commercial ice used in these tests was weakened by the significant amounts of entrapped air that it contained. In addition, the drill was allowed to entrain air, increasing its efficiency. Since the ice in an iceberg is cold and dense, actual drilling rates are expected to be smaller. On the other hand, in the more porous ice originating near the top surface of the glacier, drilling rates may be quite high, but anchor strength will be lower.
The configuration and size of the anchor were determined by the necessary holding loads. A cylindrical geometry was chosen because it provides a large surface area for anchoring and does not require complicated drilling. Handling limitations on the part of the ROV dictated that the overall tool length not exceed 15 feet. Moreover, 18 inches of unrefrigerated drill head, combined with a nearly equivalent length of anchor that cannot be completely inserted into the iceberg (near the clevis to which the tow hawser is attached), reduce the effective anchor length to about 12 feet. Therefore, the anchor's effective surface area (given a 4-inch-diameter tool) is about 1800 square inches.
Drilling creates a hole from 4 to 5 inches in diameter. The anchor is installed inside it at a depth of 50-100 feet below the sea surface. For the anchor to freeze in place, a layer of ice averaging 0.5 inch in thickness must be produced over its entire length.
Liquid carbon dioxide is available in bottles pressurized to 500 psi. With a viscosity only one-tenth that of water, it can easily be supplied in quantity through a long hose. The liquid is allowed to vaporize through orifices inside the anchor. At depths of 50-100 feet, the hydrostatic pressure is just below the triple point of carbon dioxide. Consequently, below the triple-point temperature of - 56.6 degrees C, the material flowing out of the orifices will be a mixture of solid and gaseous carbon dioxide in equilibrium. This gas/solid mixture moves through an annulus inside the anchor tube. The tube is quickly reduced to a temperature sufficiently low to maintain the equilibrium of the phases at the ambient pressure.
Laboratory tests were conducted by vaporizing liquid carbon dioxide inside a prototype anchor tube held submerged in a tank at room temperature. After 10 minutes, 0.75 inch of ice had formed. During this time, four bottles of liquid carbon dioxide were vaporized inside the anchor. The ice formed primarily near the orifice sites on the anchor and not at all near the vents, because of the turbulence induced by the large volume of gas venting. Inside an actual iceberg, the ice layer would be expected to form more quickly and uniformly, since the gas would be vented to the outside and water turbulence on the surface of the anchor would be eliminated.
For continuous towing, the desired towing load is 50 tons; for intermittent dynamic loads, with a 10-second rise time corresponding to the typical swell period, the desired towing load is 100 tons. These loads correspond to shear stresses on the anchor surface of 56 psi and 112 psi, respectively.
The shear and adhesive strengths that have been established for ice sandwiched between stainless steel plates provide a reasonable approximation of the short-term ice strength expected under dynamic loading conditions at load rates of about 15 psi/second (truly dynamic load rates, such as are experienced during impact loading, are not likely in this context). Below 13 degrees C, the shear strength is about 230 psi. At higher temperatures, failure has been shown to result from a loss of adhesion between ice and steel. In the drill-and-anchor tool, however, buttress threads provide a mechanical lock between ice and anchor. The anchor's shear strength can therefore be expected to remain near that of the ice, even at temperatures above - 13 degrees C.
The temperature near the iceberg surface where the anchor is lodged can fall between 0 degrees and - 30 degrees C, depending on the iceberg's recent history. During anchor freezing, new ice is formed at even lower temperatures. Since ice is an excellent insulator, the temperature of the ice at the surface of the anchor should remain depressed for days. Under these conditions, the dynamic anchor holding strength should equal the theoretical shear strength at low temperatures; the corresponding load is 200 tons, or twice the design load.
The area around the top of a tabular iceberg is usually composed of highly permeable snow ice originating from the upper portions of the glacier. The shear strength of this type of ice is considerably lower than that of compact glacial ice. When snow ice is present in this region, the ROV's ability to venture beneath the surface for anchoring is a valuable asset. In other cases, rolling of the iceberg can cause the zone of permeable ice to be submerged, so that saturation with sea water further weakens it. During drilling operations, it is important that an attempt be made to identify this region and to avoid it. If a weak section is drilled, the drilling rate will be high, which can serve as a warning that the anchor strength may be lower than expected.
During loading, it is possible for a radial crack to be initiated at the bottom of the anchor. Although propagation of such a fracture could cause the anchor to fail, a first-order analysis has shown that this is unlikely. In cases where anchor insertion is not complete, however, fracture may be an important failure mechanism. If the anchor's displacement due to creep in the ice during towing becomes larger than 10 percent of the thread spacing (0.25 inch), a significant reduction in the effective anchoring surface can occur, leading to failure. Analysis of creep rates in ice has shown that significant creep deformation will not occur during a 10-hour tow of a 50-ton load.
Two tests performed on icebergs in Bonavista Bay on the eastern coast of Newfoundland employed a prototype tool 3.5 inches in diameter and 12 feet long. An ROV was equipped with additional buoyancy tanks and with a hydraulic chain-drive system to provide a constant thrust on the drill of about 50 pounds. Difficulties were experienced in holding the vehicle steady at the ice face and controlling the drill feed thrust. As a result, drilling rates of only 4-6 inches per minute were achieved and complete tool penetration did not occur. In both tests, excessive vehicle motion caused drilling to be terminated after less than 30 minutes.
When the tool was released from the vehicle, the drill hydraulic lines were cut and liquid carbon dioxide was fed to the anchor. Four hundred pounds of carbon dioxide were used, enough to ensure a continuous flow through the drill for 60 minutes. Once the ROV was recovered, the polypropylene tow hawser was connected to a steel tow cable, which was let out until the workboat was 2000 feet from the iceberg. A tensiometer mounted on the cable was used to obtain accurate tow loads.
In the first test, the drill hole was 6 feet deep with about 4 feet thread engaged. The load was brought up to 30 tons for 75 minutes before failure occurred. Upon failure, the anchor was pulled out of the iceberg and dropped to the bottom of Bonavista Bay. When the tool was recovered, it was found to have bent to a 45-degree angle near the midpoint, corresponding to the length of anchor embedded in the iceberg.
The second test was more successful. The hole was drilled to a depth of 9 feet, with 7 feet of anchor thread engaged, and the load was brought up to a steady value of 25 tons for two hours. The load was then increased slowly over the next 50 minutes until failure occurred at 60 tons, corresponding to a shear stress of 130 psi. This anchor, too, was found to be bent upon recovery.
Both anchors failed at shear loads of about half the shear strength of ice. During extended loading, however, creep failure is expected. At the loads obtained, the expected displacement rate is about 0.01 inch per hour. As creep takes place, the area of ice supporting the shear will be reduced, accelerating the creep and eventually leading to failure. The anchors were also subjected to severe bending, caused by the length remaining outside the hole. These stresses caused part of the anchor thread near the iceberg surface to disengage, reducing the holding strength. Nevertheless, the pull tests performed on these partially inserted anchors showed that a full-scale anchor would be capable of providing the strength required to withstand operational towing loads.