OTC 22121 Ice Class AUV Development Art Kleiner & Jami Cheramie, C & C Technologies, Inc.; Jeremy Dean & Robert Raye, Shell Copyright 2011, Offshore Technology Conference This paper was prepared for presentation at the Arctic Technology Conference held in Houston, Texas, USA, 7 9 February 2011. This paper was selected for presentation by an ATC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright. Abstract C & C Technologies, Inc. of Lafayette, Louisiana is spearheading a technical research program to advance the state of the art regarding under ice Autonomous Underwater Vehicle (AUV) survey operations. This program, funded by Shell, focuses on five areas of research, which include: 1) Remote Stations for AUV Navigation and Communication, 2) AUV Recovery by Net, 3) AUV Recovery by ROV, 4) AUV Upward Looking Multibeam Sonar, and 5) Collision Avoidance. Form factors for the remote stations were developed following the evaluation of the potential hardware components and an examination of AUV mission requirements. Analysis was performed to determine how best to utilize the remote stations to gain the maximum utility of an AUV under the ice. Two methods for facilitating the recovery of an AUV under ice were examined: 1) recovery by a deployed net and 2) recovery by an ROV. Net recovery research involved the construction of a full-scale prototype to deploy through a hole in the ice, extend, and capture the AUV. Research into ROV recovery under ice included developing a methodology to hold the AUV in place so it could be captured by the ROV. This involved the evaluation of several scenarios, which included securing the AUV to the sea floor with a drop anchor and making the AUV negatively buoyant with either a floodable ballast tank or by the release of syntactic foam. Research into the integration of an upward looking multibeam involved the analysis of currently available systems. A comparative analysis of power consumption, size, weight, and software interfacing of the potential candidate systems was performed. C & C collaborated with Kongsberg Maritime to enhance the HUGIN AUV s collision avoidance system for under ice navigation and for marine mammal detection. This paper provides details on each of these five areas of research and details the scope, methodology, findings, and resulting recommendations. Concept Autonomous Underwater Vehicles (AUVs) are common and effective tools for seabed surveys and subsea surveillance, with benefits over conventional tow-fish systems that include reduced operational footprint, independence from weather and sea conditions, and better data quality. Their use in the Arctic is challenged by the presence of sea ice, which requires technological developments in the key areas of navigation and communications, through-ice launch and recovery, and collision avoidance. In 2009 Shell contracted C & C Technologies, Inc. to evaluate the feasibility of adapting AUV technology to Arctic under-ice survey operations, including sidescan sonar, multi-beam swath bathymetry, and sub-bottom profiling. The objective is to develop technology to allow operations within the Alaskan Beaufort Sea and Chukchi Sea during periods of ice cover, leveraging the benefits of the AUV, while extending the operational window beyond the open water season. Capabilities would include supplying positioning updates and communications to the AUV while operating under sea ice, retrieving the vehicle through ice (both disabled and under direct control), collision avoidance, and integration of upward looking sonar to measure ice thickness, study its morphology, and underwater surfaces. These enabling technologies are not currently integrated into AUVs that are commonly used for commercial AUV operations. Remote Station Research The purpose of remote stations is to reduce the risk of losing the AUV and improve mission performance. Remote stations can provide AUV communications, positioning updates, command and control, or any combination thereof (Figure 1). The cost, complexity, and size of each remote station will vary depending upon the capabilities required.
2 OTC 22121 The least sophisticated remote station is a listen only station. It would passively monitor AUV status broadcasted automatically at regular intervals, only when interrogated, or in combination. This remote station would have an Acoustic Data Link (ADL) to receive the AUV data that would then be converted to RS-232 serial data. The serial data would be fed into a data telemetry radio and transmitted to the AUV Base of Operations. The most sophisticated remote station would also incorporate an Acoustic Command Link (ACL), a DGPS, an acoustic positioning beacon, a data telemetry radio, and a Central Processing Unit (CPU). It would provide bi-directional AUV communications, real time AUV mission command and control, and AUV navigation system updates. Implementation and Station Power To minimize the risk of losing or damaging the ADL, ACL, and acoustic positioning beacon, C & C proposes to separate the transducers into a vertical array that could be lowered through the ice (Figure 2). If the transducers were unrecoverable or damaged, the electronics would remain safe within the reference station housed above the ice. The ice hole would be lined with a Telflon coated casing and embedded with a heating element. After deployment of the transducers, the water in the casing would rapidly freeze with the transducer cables in place. When removing the reference station, a portable generator would power a heater to release the casing from the ice and the ice plug from the transducer cables. A goal is to operate each remote station for three days in continuous active mode without external power generation at an internal temperature of 20C. It was calculated that 2,160 watt-hours would be required. A number of systems were researched, which included primary (non rechargeable) batteries and secondary (rechargeable) batteries (Table 1). Lead acid batteries are used in Antarctica with great success. However, at 40 degrees centigrade a sealed lead acid battery has only about 2% of its rated capacity. In addition, the weight of lead acid batteries renders them impractical for a portable station design. Calculations reveal that 592 pounds of lead acid batteries would be required to power each remote station under the frigid conditions of the Arctic. Lithium Nanophosphate batteries are rechargeable, and because of their low internal resistance, are not prone to thermal overload as is the case with many lithium-based batteries. This low resistance helps to provide better performance at temperatures below 20 centigrade. The energy per pound is not as high as the alkaline batteries, but much higher than lead acid batteries. Alkaline batteries handle low temperatures better than the lead acid batteries and are much lighter in weight. Although the capacity of these batteries is high, they are primary non-rechargeable batteries and replacing them regularly would be costly and increase personnel risk. Lithium Thionyl Chloride batteries are primary or non-rechargeable batteries. These batteries have a very high energy density and an operating temperature range between 60 centigrade to + 85 centigrade that makes them very desirable for use in Artic environments. Since they are not rechargeable, the cost of replacing the batteries would be relatively high. Generators are efficient for high power loads and are suited for extreme cold weather applications. However, maintaining and fueling several generators on a twenty-four hours per day basis, coupled with risks associated with fuel spills, potential noise contamination, and the low power requirement eliminated generators from further consideration. Solar panels are the preferred power source for many scientific stations in the Artic and Antarctic. They are efficient at low temperatures, durable, portable, and cost efficient. However, solar panels have limited effectiveness during the dark dead of winter. Ocean current turbines were considered for charging the remote station batteries. They were ruled out because ocean currents in the proposed work areas are minimal and the process of installing and removing ocean turbines through several meters of ice would be a daunting task. Wind turbines have been used to temperatures below 60 degrees centigrade and can easily meet the remote stations power requirements. Although bulky, horizontal shaft turbines are lightweight and are very efficient in most wind speeds. Vertical shaft turbines are heavy but rugged and can handle a wide range of wind speeds, including winds over 100 mph. The solution to powering the remote stations will require a combination of these technologies. Our research concludes that each remote station should incorporate a lightweight, low temperature, rechargeable Lithium Nanophosphate battery in tandem with a very low temperature Lithium Thionyl Chloride primary battery to supply as needed emergency heating. A wind turbine should supplement the battery power.
OTC 22121 3 Operating with an expected minimum current draw of less than fifteen watts, such a power configuration should operate a remote station in listen only mode for nineteen days. In standby mode, the station should operate for seventy-four days without an external power-generating source. With sufficient wind power the station could operate indefinitely. Remote Station Enclosure A protective case will be required to shield the sensitive electronics and power sources from the harsh Arctic environment and to keep the batteries at an efficient operating temperature. It would be lined with Areogel and Styrofoam insulation to hold the heat produced by the batteries and electronics. Styrofoam is inexpensive and easy to work with, but it will require several inches to insulate the electronics from the extreme cold. Aerogel fiber fabric has a much higher insulation value that allows for a more manageable thickness. Communication Testing Two data telemetry radios were tested, the Teledesign and inet. Both radios have similar communication ranges, which is basically line of sight. The tests were performed to determine if the lack of RS-232 hand shaking lines (RST, CTS, DTR...) would have an effect on the configuration or operation of the acoustic communication links. The tests were also performed to determine if the serial data transfer protocols used by the acoustic links were compatible with the serial data transfer protocols used by the radio modems. Initial bench tests demonstrated that the acoustic links could be configured through a radio link with no communication or latency issues. Also demonstrated was the compatibility of the serial communication protocols and that data could be sent from one computer, through a radio link, through an acoustic link, and to another computer with no corrupted data. System field tests took place in the Gulf of Mexico during HUGIN 1000 AUV sea trial preparations for a Shell Ice Gouge Survey program. For this test, the data radio was interfaced between the acoustic modem and the command/control computer to ensure that all AUV communications would be routed through the radio link. This test was performed with real world data during actual HUGIN 1000 AUV dives. The Teledesign radios were tested for twenty-five hours during full communication with the HUGIN 1000 AUV. The inet radios were tested for ten hours. Radio idle time was estimated at 70 to 80%, which should be ample to allow for bidirectional communication between multiple devices. These tests proved that it is possible to monitor, command, and control an AUV via remote radio control Net Recovery Research The method envisioned to retrieve an AUV from under the ice is based upon a well-proven launch and retrieval system currently employed worldwide by commercial and military organizations. The HUGIN AUV launch and retrieval system is normally positioned at the stern of a ship. It is comprised of an articulating sled mechanism that can be extended downward into the water. When the mission is complete, the AUV nose is released and floats apart from the vehicle tethered by a lightweight painter line in tandem with a heavier retrieval rope. A grappling hook is used to capture the painter line. The retrieval rope is then secured to a hydraulic winch mounted at the top of the launch and retrieval system. The AUV is then winched onto the extended sled, which is retracted safely into a horizontal position. A prototype net recovery system was designed and constructed to evaluate the feasibility of capturing the AUV nose cone from beneath the ice. Two five-foot sections of three-quarter inch stainless steel square tubing, each attached to a ring gear, act as arms to hold and unfurl the recovery net. Approximately seven rotations of a hand crank are required to fully extend the arms from a vertical to a horizontal position. A hole is positioned at the top plate to lock the handle in place when the arms are fully extended. A recovery net was selected with a square mesh of three inch spacing and a total surface area of 100 square feet. The net is attached to the square stainless steel tubing at four points using office binder type clips providing a holding force of approximately four to seven pounds. A weight is attached at the bottom of the net to keep it extended. Approximately fifteen pounds of force is required to separate the net from the frame. The AUV nose will be modified to incorporate lightweight exposed barbs to permit it to entangle in the netting. The modification proposed would include two rows of perpendicular capture rods constructed from stainless steel round bar with a diameter of approximately six millimeters. One row would be inclined horizontally toward the forward motion of the vehicle and the other would be inclined in a vertical orientation. The rods would protrude from the body of the AUV nose approximately sixty millimeters. Threaded inserts will facilitate simple and expedient replacement.
4 OTC 22121 AUV Net Recovery Procedure During normal retrieval operations, the AUV will be commanded to circle at a programmed depth near the recovery site. When AUV control is established and tracking is steady, the net recovery system will be deployed to a safe depth beneath the ice canopy. The AUV operator will command an appropriate vehicle heading, speed, depth, and lead in distance. When the AUV entangles in the netting, the net will separate from the extended arms of the recovery system. At this time, the AUV operator will cease vehicle forward propulsion and command release of the AUV nose cone. The recovery arms will then be retracted and the net recovery mechanism removed from the ice hole. The net, with the AUV nose entangled, will be pulled upward though the ice hole by an additional painter line, which was secured to the net prior to its deployment. The AUV retrieval rope will then be captured and secured to the launch and retrieval winching system. ROV Recovery Research Should the AUV nose cone not become entangled during the net recovery procedure, a Remotely Operated Vehicle (ROV) would be deployed to capture the AUV nose. To accomplish this, the AUV would need to undergo a change in buoyancy and remain stationary anchored on the sea bottom. Several methods were evaluated including releasing an anchor, releasing foam floatation, employing a floodable ballast tank, and employing a releasing air ballast system. The following criteria were considered critical to the operation: 1. Must allow the AUV to be held to the seafloor by twenty-five to forty pounds of force. 2. Must be reversible for recovery. 3. Must be safe to perform. 4. Must be environmentally friendly. Anchoring System Evaluation The AUV anchor system would include a weight, rope or cable, and reel mechanism. Developing such a system for the HUGIN could be as simple as a few modifications to the existing drop weight that is currently incorporated on the AUV. When deployed, the weight should hold the AUV suspended in the water column and keep it from drifting. In this scenario, an ROV would disconnect or sever the tether line between the anchor weight and the AUV. Since additional ROV missions would be required to locate and retrieve the weight from the ocean bottom, this technique was deemed undesirable. Foam Release Evaluation In this concept, the AUV releases floatation, becomes negatively buoyant, and sinks to the sea floor. The released flotation would have to either be recovered or be secured to the AUV by a tether greater in length than the overall water depth. This method was deemed risky and also undesirable. Ballast Tank Evaluation This concept incorporates a pressure container with a volume of air displacing between fifteen to twenty-five pounds of water at working depth. During normal and emergency decent, solenoid valves would allow seawater to enter the pressure container. High-pressure air would be employed to force the water out of the pressure container through a low-pressure check valve making the AUV become positively buoyant and rise to the surface. In addition to hardware and software design and development, this system would require extensive AUV body modifications in conjunction with extensive performance and safety testing. For these reasons, in addition to the fact that the pressure container would have to be quire large to be effective, the ballast tank concept was rejected. Releasable Air Ballast Evaluation A collapsible air bladder ballast system was also considered. This would require two collapsible air bladders (distributed on each side of the vehicle for ballasting purposes) capable of holding a greater pressure at the surface than at maximum operating depth. When at depth, a valve would release air to provide negative buoyancy. Research determined that commercial off the shelf lay-flat hose could be utilized for this purpose. These industrial hoses are designed to work in harsh environments (including at low temperatures) and to withstand high working pressures. End caps would be designed, machined, and tested to maintain pressure in the hoses (Figures 3 and 4). Forward end caps would mount directly to the AUV and accept filling and releasing lines. Rearward end caps would mount in a bracket to allow them to slide back and forth when compressed and pressurized. This concept appears to be the most practical approach to ballasting the AUV for ROV recovery under ice. Four feet of three-inch diameter hose per side would create a neutrally buoyant system with an anchor force capable of eight times the weight of a single end cap (Tables 2 and 3).
OTC 22121 5 Upward Looking Multibeam Sonar Research Part of this research was to evaluate the integration of a multibeam sonar in the AUV in an upward mounted orientation to examine the bottom of the ice cover. Multibeam data could be used to study the ice morphology including keels, ridges, leads, texture, and thickness. Sonar Requirements The first task of this research was to evaluate commercial multibeam sonar systems that would provide appropriate data and integrate with the AUV, a Kongsberg HUGIN 1000. When integrating any sensor on an AUV, the two dominant, first tier, constraints are the power and space required. A list of second tier requirements includes multibeam swath angle, beam size, sonar frequency, slant range, and autonomy of operation. Peripheral requirements, such as computer and operating system support and the maximum depth rating of the sonar transducers, were also considered. Swath angle, beam size, sonar frequency, and slant range are important in determining the suitability of the sonar to map the underneath of the ice, while autonomy, support PC, and OS are limitations of the AUV platform. Autonomy of the multibeam sonar can be defined as its capability to collect and process data with little or no human intervention. This is important on an AUV due to the limited bandwidth available for command and control. The sonar must be able to track and detect with minimum user intervention. It should also be able to auto range and auto gain to maintain bottom detection in a changing environment. Given the layout and space limitations, the need for a secondary support computer to control the sonar could be problematic. It was determined that should a secondary computer was needed to control the sonar, the AUV payload-processing computer could mitigate this requirement. The Culling Process Three multibeam sonars from various vendors were examined as candidates: 1) Odom Hydrographic System ES3 2) BlueView Technologies MB1350 3) Tritech International Limited Horizon Multibeam Echosounder System From the size and power requirements, it was determined that the Tritech sonar transducer was too large to mount in a HUGIN 1000 vehicle (Table 4 and Figure 5). It was determined that both the Odom and BlueView systems were viable options and the research progressed to compare these two systems from a hydrographic perspective. After analyzing the specifications of both systems, it was deemed that the Odom ES3 was the better choice due its lower frequency and greater achievable swath width (Table 5). Although the BlueView system was capable of matching the swath coverage provided by the Odom ES3 system, a dual-head transducer configuration would be required. This would increase the cost and complexity of the integration. An evaluation was conducted to establish if the Odom system was capable of meeting the second tier requirements relating to autonomy, computer, and operating system support. Following extensive dialog with Odom technical personnel, it was determined that the Odom ES3 was the best candidate multibeam system to integrate into a HUGIN 1000 AUV. Multibeam Integration Research A HUGIN vehicle assessment revealed that space was available in the upper aft half of the mid section with dimensions of approximately eighteen inches in length by ten inches in width by eight inches in depth (Figures 5 7). Given the dimensions of the Odom multibeam sonar, it was decided that this was the best location to mount the transducer in an upward looking orientation approximately six feet rearward from the tip of the AUV nose. The Odom multibeam sonar is composed of a processing unit and transducer in one self-contained unit. Interface for power and communications to the processing unit can be accomplished via a single cable incorporating an eight-conductor sub sea connection. The sonar operates on twenty-four VDC nominal with a range of nine to thirty VDC with expected total power dissipation less than twenty-five watts. Communication with the multibeam sonar-processing unit is via Ethernet utilizing two pairs of twisted wires to provide a ten base-t Ethernet connection. The integration of the upward looking multibeam sonar into the normal data collection and post-processing routines of the AUV require two major software development tasks: 1) Command, control, and logging of data to and from the multibeam sonar and 2) Real-time and post mission processing of the multibeam sonar data. Command, control, and logging of the multibeam data requires writing a controller to communicate with and configure the multibeam sonar-processing unit via operator chosen settings. This development would also allow the AUV to start the multibeam sonar in a valid initial state, begin pinging, and collect valid multibeam sonar data without human intervention.
6 OTC 22121 Further, it would provide a bi-directional interface for transferring commands to configure the multibeam sonar unit via remote user interfaces. The data processing and logging subsection development would ensure that all raw data is logged in the appropriate format and stored according to the prescribed line names defined prior to each mission. It would also provide feedbacks of any irregularities in the data stream collected by the multibeam sonar. Real-time would provide waterfall displays and other feedbacks to the AUV operator for quality control and analysis in addition to a user interface to the multibeam sonarprocessing unit. Post mission will include all decoding and geo-referencing of the multibeam sonar data for incorporation into software tools used to generate maps and other products for environmental analysis and delivery to clients. Collision Avoidance Research Operating an AUV under sea ice presents a hazard from deep pressure or shear ridge ice keel and grounded keels, which are common in shallower water coastal areas. Additionally, Shell has specified that the AUV shall be launched and recovered through fast ice. This requires the AUV to navigate in conditions of low overhead and through the shear zone, where threats from ice keels are more likely. The Arctic-Class HUGIN 1000, developed by Kongsberg contains a sensor system that includes a wide angle of view mulitbeam sonar that scans ±60 in the vertical plane and ±10 horizontally in the direction of travel. In addition, the system includes two narrow-beam altimeters looking forward and upward for greater detection range and redundancy. Five scenarios were tested in simulation: rugged terrain, underwater hill, floating obstruction, blocked passage, and midcolumn target. The vehicle successfully responded to the simulation challenges by changing depth, circling around for different approach vector, or returning to a rendezvous point per rules executed by the Collision Avoidance supervisory algorithm. For through ice missions anticipated in the near term, the vehicle s utilization will be limited by the complexity of the under ice environment. There is high confidence that the vehicle will successfully perform survey missions under simple ice cover. However, grounded or deep keels that are sufficiently broad or numerous will likely trigger the vehicle to abort. Use of the vehicle in areas or conditions of pronounced shear/compression or suspected groundings is not contemplated, since it is beyond the capabilities of the present hardware and software. A promising application of the collision avoidance system is to detect the presence of marine mammals in the path of the vehicle. A detailed plan of action was developed by Kongsberg to modify the collision avoidance system and system architecture currently embedded within the HUGIN 1000 to detect and track obstructions moving through the water column. The scope of this development involves software implementation, simulation, and sea trial verification. Implementation includes assessing marine mammal avoidance solutions, software development and design, and iterations to ascertain performance. Simulations would be performed for a variety of scenarios to verify performance and to quantify and document system response attributes. Kongsberg and C & C Technologies will accomplish sea trial verification jointly. A sonar reflector simulating an underwater mammal will be constructed and performance tested with the Forward Looking Sonar. Initial testing will be performed in a test tank followed by open sea experimentation with a HUGIN 1000 vehicle. Ultimate verification of this enhanced capability will require testing under real world conditions in actual Arctic AUV missions. Conclusions This research program was performed to evaluate the feasibility of adapting conventional AUV technology for Arctic under ice seafloor survey operations and expand operational capabilities specific to the Alaskan Beaufort Sea and Chukchi Sea. The technology areas researched are not presently implemented for commercial AUV operations. Initial research results indicate that the technology is obtainable and components are available, which can be merged and configured to expand AUV mission capabilities for under ice applications. Once implemented, the benefits for Arctic applications include; reduction in marine traffic and noise, extended operational season, detection of marine mammals, enhanced data quality, and ice and weather independence. Going forward, Shell will be looking for opportunities to field trial the modified HUGIN 1000 and supporting technologies in the Beaufort Sea during the late winter/early spring from fast ice. This includes preparatory development and field testing of the remote station, through ice recovery, and upward looking sonar concepts. Acknowledgements The authors with to thank Robert Henry, David Leger, and Peter Alleman for the concepts and technical research upon which this paper is based. We additionally wish to recognize the significant contributions of Kongsberg Maritime and the University NAVSTAR Consortium (UNAVCO).
OTC 22121 7 Figure 1 Remote Station Concept for AUV Under Ice Surveys Figure 2 Remote Station Communications/Navigation Hardware
8 OTC 22121 Figures 3 and 4 End Cap Design for Releasable Air Ballast System Table 1 Remote Station Battery Comparison Battery (Type) Sealed Lead Acid (Secondary) Lithium Nanophosphate (Secondary) Alkaline (Primary) Lithium Thionyl Chloride (Primary) Low Temperature Performance Weight Relative Cost Poor (15% power at -20C) Very Heavy (610 lbs.) Low Fair (60% power at 20C) Heavy (73 lbs.) High Poor (stops working at 20C) Light (for emergency heating only) Low Excellent (rated to 60C) Light (for emergency heating only) Low Table 2 Lay-flat Hose Buoyancy Calculations Diameter in Inches 3 4 5 Lift per Foot in Pounds 3.00 5.48 8.60 Table 3 - End Cap Weight Calculation Weight of Three- Inch Diameter End Cap Item Description Hose Cap Guide Clamp Ring Total Volume in Cubic Inches 10.04 3.61 4.47 Weight in Pounds 2.9 1.05 1.29 5.25 Table 4 - Multibeam Sonar Size and Power Comparisons System Dimensions Power Consumption Weight in Air Weight in Water ES3 162mm (6.3 in.) L 117mm (4.62 in.) H < 25 watts 8 kg (18lbs.) 1 kg (2.5lbs.) 92mm (3.63 in.) W MB 1350 10.5 x 9.2 x 5.3 in 15 watts 12 lbs. -3.2 lbs Horizon 467mm (18.39 in.) W 334mm (13.15 in.) H 116mm (4.57 in.) D 60W 20Kg (44lb) 10Kg (22lb)
OTC 22121 9 Table 5 - Multibeam Sonar Technical Specifications System Frequency Swath Width Number of Beams Max Range/ Slant Range Max Ping Rate ES3 240 khz 120 deg 480, 240 or 120 60 m 100 m MB 1350 1.35 MHz 45 deg 256 25 m 30 m 14 Hz 40 Hz Figure 5 Proposed Mounting Location (in yellow) for Odom ES3 Multibeam Sonar Figures 6 and 7 Proposed Mounting Location of Odom ES3 Multibeam Sonar