CINEMar/Open Ocean Aquaculture Annual Progress Report for the period 1/01/04 through 12/31/04

Principal Investigator: B. Celikkol, K. Baldwin, D.W. Fredriksson, M.R. Swift, I. Tsukrov

I. Accomplishments

A. Scheduled Tasks
Background: For the past six years, the University of New Hampshire (UNH) has operated an open ocean aquaculture site in 52 meters of water approximately 10 km from the New Hampshire coast in the United States. The site is permitted to perform research related to the operational, engineering, biological and environmental aspects of open ocean aquaculture (OOA). To support the research, two independent 600 m³ Sea Station™ fish cages (SS600) were analyzed and deployed at the site in 1999 using separate, robust mooring systems (see Tsukrov et al., 2000; Fredriksson et al., 2000; Baldwin et al., 2000). For over four years, these systems were the focus of an intense engineering and operational analysis program. From the engineering perspective, studies were conducted to investigate the dynamics so that numerical and physical modeling techniques could be developed to cost-effectively engineer and specify equipment suitable for deployment (see Palczynski, 2000; Tsukrov et al., 2003; Fredriksson et al., 2003a; 2003b).

Last year, an effort was made to expand bio-mass capacity at the site; the two small systems were replaced with a larger four grid mooring enabling the deployment of additional containment structures (Fredriksson et al., 2004). The new mooring system also allows auxiliary equipment, such as feeding platforms (Rice et al., 2003; Fullerton et al., 2004), to be installed at the site. The intent is to approach commercial level operations so that proper economic assessments can be initiated. Even though the techniques and engineering methods are being developed, engineering challenges still have to be overcome to create the technological basis for commercial scale aquaculture.

Objectives: To address these challenges the goal of the OOA project has been set to establish a commercial scale facility at the Isles of Shoals site. The engineering component supports this collective goal by pursuing the following objectives,

1. Investigation of commercial scale fish cage systems.
2. Analysis of the existing mooring grid system at the OOA site.
3. Feed buoy development including design and control/telemetry.
4. Development of an operations center.
5. Improved numerical modeling techniques.

These objectives represent a continuation of previous work performed (see Baldwin et al., 2003 or the OOA Engineering progress report for 2003). In the following section, a brief description of each of these studies is included.

B. Progress on Tasks
PROGRESS ON TASK #1: INVESTIGATION OF COMMERCIAL SIZE SYSTEMS
As part of the proposed work, the Center for Ocean Engineering (COE) at UNH has been involved in the analysis of possible commercial size fish containment systems. JPS Industries, a NH manufacturing company, worked with UNH to obtain a Small Business Innovation Research (SBIR) project. The Phase I technical objective is to develop a low-cost, submersible fish cage for open ocean aquaculture. The approach includes the following tasks to be completed over the six month period, August 2004 through January 2005.

1. Develop a conceptual design
2. Conduct hydrostatic analysis
3. Perform numerical modeling of systems
4. Perform physical modeling of systems
5. Perform structural modeling of systems
6. Specification of components for Phase II
7. Phase II proposal development

The first task to be completed in this project is to develop a conceptual design(s) so that hydrostatic analysis can be performed. At this stage, the goal is to design a cage with a containment volume of approximately 5000 m³. Two concepts have been investigated and are described below.

Concept 1: This cage concept (Figure 1a) is a modified version of a traditional surface gravity cage used by many salmon farms in Maine and Canada. It consists of 2 rings of High Density Polyethylene (HDPE), held together with stanchions. A handrail is located on top of the stanchions. The bottom ring typically consists of one or two rings, filled with sand or concrete. The intent is to modify the traditional cage to be submersible. To do this, the two upper cage rings will be sealed airtight or filled with individual flotation units to provide buoyancy to the system. A bridle, made either of chain or line, will be configured to connect the lower rim(s) to an airlift. Below the airlift will hang ballast weight (which can consist of chain or deadweight anchor(s) depending upon bottom conditions). The variable ballast (to raise and lower the cage) is located in the airlift.

Concept 2: A second design, shown on Figure 1b, is also being considered. This design moves the variable ballast from below the cage to the top superstructure. This modified cage design consists of one ring of HDPE (sealed or filled with flotation), which has modified stanchions that run horizontally to an inner rim (approximately _ of the cage diameter). This rim (and one above it) forms a cylinder. This can allow the majority of the cage to be submerged, yet have access at the surface via the inner top rim (if needed). The variable ballast (to raise or lower the cage) will be contained in the vertical supports of the inner cylinder. The bottom rim consists of the same superstructure. Sand or concrete can fill the outside ring and chain will hang off various sections of the rim.

The next task to be completed consists of performing hydrostatic calculations for each of the geometric configurations for the design concepts. The calculations are not included in this progress report (for brevity) because they consist of approximately 24 worksheets.

During the initial design process, it became apparent that Task #5, “to perform structural modeling of the systems” was a critical component in the decision to choose a prototype. Therefore work on this task was initiated. The work included construction of a cage upper rim model for concept #1 (see Figure 2). One quarter of the cage rim, handrail, and associated stanchions was created using 16,466 elements and 16,375 nodes using the structural model Marc/MENTAT. The model was built using material properties characteristic of typically gravity cages. When a load is applied in the structural model, stresses are calculated to determine if the structure can withstand a specified environmental condition. The load simulated the force applied by a mooring bridle line, which field experience has shown to cause buckling failure.

PROGRESS ON TASK #2: EXPANSION OF MOORING SYSTEM GRID
Background: In the last year, the COE team designed and analyzed a four-grid, submerged mooring system (Fredriksson et al., 2004). In the summer of 2003, the operations team deployed it. Over the last year, many new cage and mooring system options have been discussed. To investigate the feasibility of some of the new “proposed” concepts, additional model simulations were conducted to assess the holding capacity of the grid mooring. Calculations were completed using the UNH finite element computer program Aqua-FE

UNH 4-Cage Grid System with a Single Gravity Cage: One of the first scenarios to be considered includes an investigation of the 4 cage grid system with the quarter-ton and one-ton capacity feed buoys, one 3000 m3 Sea Station (SS3000), one 600 m³ Sea Station (SS600), and one 5000 m³ experimental cage (similar to the concepts described as part of Task #1 called the “Yankee”). A schematic is shown on Figure 3. Simulations were performed using the present grid configuration. In the model, the waves and currents approached the system from the Northeast. A 9 meter wave with an 8.8 second period and a 1 m/s current was used as input. Results of the numerical model simulation focused on the six anchor legs deployed in the north and east direction (Figure 4). The grid lines were also analyzed because these components play a role in the distribution of loads through the system. In the model, the net chamber depth of the gravity cage is 11 meters (36 feet). The net on the gravity cage has a twine diameter of 2 mm, a stretch length of 48 mm (1 7/8”). The pennant weights on the 2 Sea Station cages were modeled as fixed points.

The load distribution throughout the mooring lines is shown on Figure 5. Keeping in mind that the nominal design load for the system is 177 kN (40,000 lbf), anchor line 2 exceeds this amount by 2.7 percent. It was calculated to have a load of 182 kN (41,000 lbf). Anchor dragging in severe storms, could therefore, result unless the anchor capacity is upgraded.

UNH 4-Cage Grid System with two Gravity Cages: Another modeling scenario consisted of the 4 cage grid system with the quarter-ton and one-ton capacity feed buoys, one SS3000, one SS600, and two 5000 m³ experimental cages (Figure 6). In the model, the waves and currents approached the system from the Northeast. A 9 meter wave with an 8.8 second period and a 1 m/s current was used as input. Results of the numerical model simulation focused on the six anchor legs deployed in the north and east direction (Figure 7). The net chamber depth of the gravity cage is 11 meters (36 feet). The fish net on the gravity cage has a twine diameter of 2 mm, a stretch length of 48 mm (1 7/8”). The pendant weights on the 2 SS cages were fixed.

The load distribution throughout the mooring lines is shown on Figure 8. As previously mentioned, the nominal design load for the system was determined to be 177 kN (40,000 lbf). Results from this analysis indicated that anchor legs 2, 3, and 5 could possibly exceed the design load. The calculated loads using the numerical for anchor legs 2, 3, and 5 were 189.6 kN (42,600 lbf), 202.9 kN (45,600 lbf) and 215.6 kN (48,500 lbf), respectively. Upgrading grid anchoring capacity should, therefore, be seriously considered before deploying this amount of aquaculture equipment.

Load Cell Deployment: Another portion of the work being performed as part of the analysis of the four-grid mooring incorporates deployment of load cells in the Northeast corner of the grid system and comparing field results with the numerical model. As described in the proposal, it was planned to deploy the load cells in Year II of the project since the equipment was being used as part of a complementary project in Eastport, Maine at a salmon farm (see Section F for a brief description of the project). All of the loads cells returned in good order and are now being refurbished for deployment at the OOA site.

Pendant Line Load Cells: Originally, another load cell was scheduled for deployment as a component in one of the cages pendant line. The intent of this load cell was to monitor the bio-mass accumulation on the cage. However, deployment of the load cell, attachment shackles and other components would ultimately place the cage too shallow. Therefore it was decided not to deploy the instrument.

PROGRESS ON TASK #3: FEEDING TECHNOLOGIES DEVELOPMENT
The following sections were included as part of progress on this specific task:

1. One-ton buoy mooring,
2. One-ton buoy electrical engineering,
3. Feed buoy SBIR Phase 1,
4. Feed buoy SBIR Phase 2 proposal/cooperative agreement and
5. Twenty-ton feed buoy design.

One-Ton Feed Buoy Mooring: The newly constructed one-ton feed buoy was towed to the OOA site on December 10, 2003 and moored in the vacant grid bay as shown in Figure 9. This “temporary mooring” consisted of two low-stretch bridle lines having mid-span submerged weights to maintain tautness. The slack feed hose was loosely lashed along bridle lines to the cod cage and was not structurally part of the mooring system. This was discussed, in part, in the report submitted for 2003, but it is included here for background purposes.

A new mooring system was designed so that the one-ton feed buoy could be moved out of the spare grid bay thereby freeing up space for a possible additional cage. The mooring consisted of three compliant tethers constructed of specially designed, reinforced rubber hoses. The tether-hoses were secured to grid intersection points to locate the buoy over an outside grid point adjacent to the cod cage. The feed hose leading to the cod cage would be shorter and remain non-structural as far as the mooring system is concerned.

In support of the 3-point design effort, Froude-scaled wave tank studies were completed using the 1/24 scale physical model shown in Figure 10 (left). Other long-range objectives were to fully characterize the seakeeping dynamics of the one-ton design thus adding to our cumulative knowledge of feed buoy wave response. Building a data base of feed buoy dynamic characteristics is important in the future design of larger buoys.

In preparation for tank testing, the scale model was refurbished and re-ballasted to represent the fully loaded, “as-deployed” condition. The physical model was tested in UNH’s 36 meter long by 3.66 meter wide by 2.48 meter deep wave/tow tank. Experiments were carried out adjacent to a window in the side of the tank, so an Optical Positioning and Evaluation System (OPIE) could be used to measure buoy motion. As described by Michelin and Stott (1996), the OPIE system consists of a digital camera that records images at a user-set frequency, a dedicated computer with frame-grabber, and processing software programmed in MATLAB. Small black target dots are placed on the white painted model, and the model is illuminated so that the black dots stand out from the much lighter background. The software operates by tracking the black dots on each succeeding image. Each recording must be calibrated so that distances in number of pixels can be converted to conventional distance units. All images for the feed buoy tests were recorded at 30-frames/second.

Heave and pitch natural frequencies were determined from free-release experiments done on April 12, 2004 using the model buoy (McGillicuddy et al., 2004a). For heave tests, the model was raised slightly from its equilibrium position and released from rest (see Figure 10 - right). Pitch testing was similar except the model was tipped (without changing its vertical position) and released from rest. At least three replicates were done for each type of test.

Heave damped natural period was found to be 3.5 seconds (all results presented at full scale). Since this is on the short-period end of the normal wave environment (3 ­ 10 seconds), wave contouring behavior is expected for moderate to large seas. The pitch damped natural period was 8.4 seconds putting the pitch resonance condition (buoy only) in the middle of the wave energy range during storms. (Though attempts had been made while developing the design to alter pitch resonance by changing ballast and adding damping baffles, little advantage was possible, and the maximized reserve righting moment attribute already achieved was regarded as the highest priority.)

Wave testing of the buoy and 3-point compliant mooring (see Figure 11) took place on April 28, 2004 (McGillicuddy et al., 2004b). Due to tank width limitations, the full buoy mooring/grid could not be set up. Instead, the three grid point attachments for the tether-hoses were fixed at the tank walls. Essentially, it was assumed that the feed buoy dynamics would be most influenced by the directly connected, compliant members, while remote components (such as grid anchor lines) would play a diminished role. Regular wave, single frequency tests were done to obtain normalized transfer functions in the form of Response Amplitude Operators (RAOs).

The maximum heave RAO (heave amplitude divided by wave amplitude) was 1.67 and occurred at the heave damped natural period of 3.5 seconds. The maximum pitch RAO (pitch amplitude in degrees divided by wave amplitude in meters) was 11.4 degrees per meter and occurred at a period of 7.1 seconds (slightly shorter than the pitch damped natural period of 8.4 seconds). RAO results revealed no threat to the buoy’s ability to survive site conditions, and during the tests, no unusual behavior was observed visually.

Despite the advantages of the 3-point compliant mooring system, this design was not implemented due to the cost of the tether-hoses. Instead, a hybrid system was devised consisting of two low-stretch lines with central ballast weights and one tether-hose that was available from a previous project (see Figure 12). The combination of the hose, lines and weights provided an adequate restoring force to contain the buoy in a safe watch circle to prevent it or any other gear from coming in contact with the cage. The mooring geometry and system response of the mooring was tested using Aqua-FE. The system was deployed July 29-30, 2004.

One-Ton Feed Buoy Electrical Engineering: Overall, the electrical system of the one-ton feed buoy has exceeded expectations and has worked as designed, safely operating through several mechanical additions and changes. The system was designed to be flexible in terms of ability to power components as they are installed, and this flexibility has been utilized extensively. When a new 240 VAC mixing chamber wash-down pump was added, all that was required for wiring was plugging the new pump into a spare connector on one of the electrical panels. The same configuration worked well when biologists wanted to add an acoustic fish tracking system and needed a battery bank charger. All that was needed was to plug into another spare outlet.

In September of 2004 a major electrical advance took place for the one-ton buoy when underwater cage lighting was installed. Two 120 VAC metal halide lights, built by JT Electric, Faeroe Islands, were installed to prevent fish maturation by simulating extended daylight during the short winter days. The design posed several challenges in terms of safely powering high voltage underwater lighting without endangering divers that may possibly be working near the powered lights. Impulse Enterprises provided underwater pluggable connectors to suit the application and the system deployed smoothly. The two lights are being run twice daily from 5:00 to 8:00 in the morning and 5:00 to 8:00 in the evening. This was a big step towards developing a commercially viable aquaculture system since cage lighting is being utilized at almost all major aquaculture operations.

The electrical system on its own could not be as successful as it has been without the programming assistance of Jim Irish of Woods Hole Oceanographic Institution (WHOI). The UNH engineering team had to fully develop and construct every component of the electrical system from the microcontroller to the high power control relays and then convey exact instructions of how each of these components is to be controlled. This is much more complex than it first appears in the sense that this system is fully autonomous. If something goes wrong (i.e. a breaker trips, emergency stop switch is activated, generator won’t start, etc.), the control system/program has to have the ability to detect the problem and take corrective actions to remedy it. The program needs to have a certain level of logic to it to determine how to take corrective action. For instance if the breaker for one of the main feed pumps trips, the program must know that it has tripped and to stop the regular feeding sequence so that the feed hose does not form a clog. The system also reports this problem to shore, and feed will not be distributed until the problem is corrected.

New Feed Buoy: SBIR Phase I: Another major initiative, complementary to the OOA engineering project, is the development of a commercial size feeding system. The initiative is a part of a Small Business Innovative Research (SBIR) project with Net Systems from Bainbridge WA. Over the last year Phase I of the SBIR was completed and Phase II is underway. This section describes, in part, the work that was completed during Phase I (Swanson et al., 2004).

Originally, an eighty-ton capacity feed buoy design was investigated in the cooperative effort with Net Systems, Inc. Net Systems generated a working design consisting of a steel spar buoy enclosing four feed silos and machinery for feed transfer. The overall shape is illustrated in the Figure 13 photo of the physical scale model used in tank testing. The main cylinder is 4.55 m in diameter and 11.8 m tall. The lower cylinder, 1.2 m in diameter and strongly gusseted to the main cylinder, secures the concrete ballast disk ­ 1.5 m thick and 5.59 m in diameter. Fully loaded, the feed buoy has a mass of 1.89 (10⁵) kg. The main role of the UNH team was to construct the physical scale model and evaluate wave dynamic performance by means of tank testing.

The model feed buoy was built to a scale ratio of 1: 20.77 which is approximately the ratio of tank depth to depth at the UNH Open Ocean Aquaculture site, a possible deployment location. Froude-scaled tank testing began with free release experiments (no waves) in heave and pitch. Measurements were used to infer natural frequencies and periods, damping ratios, added mass and added mass moment of inertia. Next, regular wave response testing was done to obtain Response Amplitude Operators (RAOs) in heave and pitch. Wave testing was done over full scale frequencies spanning the energy containing waves at the OOA site. Waves were mostly deepwater with the longer period waves going into the intermediate wave regime.

The completed buoy model (see Figure 13) stood about 3 feet tall and weighed about 45 lbs. The upper and lower cylinders were made of PVC pipe. The base ballast disk was cast in concrete, the same material as that in the prototype design. The buoy was outfitted with internal ballast to reflect the scaled, overall weight distribution when fully loaded since this had the highest center of gravity placement.

Free release tests were conducted to determine heave and pitch natural frequencies and damping ratios. Experiments were carried out adjacent to a window in the side of the UNH wave/tow tank using the OPIE positioning system (see one-ton mooring section for description). The free release experiments were done with the buoy model only. For heave tests, the model was raised slightly from its equilibrium position and released from rest. The model oscillated vertically with decaying amplitude. Pitch testing was similar except the model was tipped (without changing its vertical position) and released from rest. At least three replicates were done for each type of test.

Heave damped natural period was found to be 8.5 seconds (all parameters full scale). This is within the likely period range of a normal wave environment (3 ­ 10 seconds). In fact, heave period is very close to the expected storm wave periods at the OOA site. The pitch damped natural period was 13.8 seconds putting the pitch resonance condition (buoy only) on the long period side of a likely wave environment

Wave response of the eighty-ton feed buoy design was characterized by conducting a series of single frequency wave tests (see Figure 14). The regular wave experiments were carried out at periods ranging from 3.75 seconds to 18.23 seconds thereby spanning the range of high energy forcing, as well as bracketing the heave and pitch damped natural periods. Two mooring configurations were used which employed two horizontal bridle lines (with and against the wave direction) attached to the base of the main cylinder. Each bridle line led horizontally to a subsurface buoy anchored to the bottom. In one case, the bridle lines were pre-stressed at 4800 lbs, and in the second case, no pre-stressing was used.

Heave response was not sensitive to the differences in the two mooring configurations employed. Both had an extreme resonant response at the heave damped natural frequency. The maximum heave RAO (heave amplitude divided by wave amplitude) was 3.2. At heave resonance, vertical motion was so extreme that out-of-plane (cross-tank) roll motion was induced. Pitch response was also large in the vicinity of the pitch damped natural frequency. Maximum pitch RAO (average between the two mooring configurations) was 34.5 degrees per meter of wave amplitude.

The large heave RAO, the match between heave resonant frequency and storm wave frequency, as well as loss of lateral stability, indicated that the tall spar configuration was not a suitable feed buoy shape. Previous UNH buoy tank test data showed, moreover, that a shorter buoy having a larger water-plane area would have a shorter heave natural period, more heave damping and a much reduced maximum heave RAO. Thus subsequent work in feed buoy design was directed towards more can-like shapes.

New Feed Buoy: SBIR Phase II: At the end of the Phase I SBIR, it was decided to redirect the feed buoy design to a smaller, more can-like shape than the eighty-ton model previously analyzed. Furthermore, projected costs to build the eighty-ton system would most likely exceed the amount available for funding through the SBIR Phase II program. Therefore, a unique collaborative agreement was made between Net Systems, the OOA operations group and OOA engineering. It was decided to supplement the needs to complete the Phase II project by using OOA operations resources necessary to acquire hardware for the system and OOA engineering personnel funds (labor) to assist in the design and analysis of the system. So far the arrangement has worked well and the design of the system has been worked on for over six months. The schedule is to complete the design early in 2005, construct the buoy at an outside facility in late 2005 and launch, deploy and test the system in early 2006. An update of the progress is provided below.

The first step in the process was to develop a strict set of design criteria. These criteria are the result of numerous meetings with all groups involved represented. A summary of the overall design criteria is provided below.

1. Feed capacity:
   • 20 tons (40,000 lb)
2. Feed storage:
   • Four separate storage bins, to allow four different species/sizes of fish to be fed from the same buoy.
   • Different species require specialized feed.
   • Varying stages of development requires different size feed
3. Hydrodynamic Characteristics
   • Survive design wave condition (9 m height, 8.8 sec period, 1 m/sec current).
   • Survive storm spectrum (e.g. H_mo = 8.25 meters, T_p = 12 seconds).
   • Positive righting moment at all angles under all conditions.
   • Withstand upsetting moments and added weight due to ice buildup.
   • Have minimal list due to unbalanced feed storage.
   • Watertight integrity
4. Machinery
   • Above waterline (whenever possible).
   • Components should be removable if repair/replacement is necessary.
5. Power system
   • Power generated for all electrical needs.
6. Control & telemetry
   • System to allow user to control feeding systems from shore station.
   • System monitoring of important electrical systems.
7. Mooring
   • Buoy outside of grid (due to grid limitations).
   • To accommodate height variations due to storm surge, tide and wave elevation.
   • To sustain storm conditions specified in Criteria 3.
8. Maintenance
   • Internal components (systems) must be accessible for routine maintenance while deployed.
   • Maintenance cycles to be established.
9. Safety systems
   • Incorporate safety features into the design. Including fire suppression systems, air quality monitoring and control, radio communication systems and abandon ship equipment.
   • Suitable escape routes to be provided.

A meeting with engineering, infrastructure and biology group members on 12 October 2004 was conducted to review and finalize the Feeding System Design Criteria (External Feed Transfer). A summary is this set of criteria is provided in bullet form below.

1. General
   • Any storage bin (silo), total of four, should be able to feed any cage, total of four.
   • Feed one cage from one silo at a time.
   • Four separate feed hoses exit the buoy.
   • Each feed hose is connected to one cage.
2. Feed amounts (flow rates)
   • A maximum feed transfer rate of 1800 lb/hr. Assuming an average feed density of 37 lb/ft³, the volume feed rate is approximately 50 ft³/hr.
   • A time limit of 5 minutes of transport time from buoy to cage
3. Geometry
   • The maximum assumed length of tube from buoy to cage is 800 ft.
   • Approximate discharge height of feed tubes from buoy is 10 ft above the waterline ­or­ the feed tubes will be exiting through the bottom of the buoy.
   • Buoy to cage pipe size diameter is either 3 or 4 inches.
4. Size
   • Space is limited in the buoy, so machinery size should be kept to a minimum.
   • Since the majority of the system will be located high in the buoy, keeping the weight low is important.
5. Other considerations
   • A continuous feeding system is desired over a batch feeding system. This is to minimize the amount of time required to feed.
   • Feed transfer rate most likely be slower than the water transfer rate due to the feed and pipe wall interaction (including turbulence of flow).
   • System should be able to survive all situations that the feed buoy will experience, including storm waves and freezing temperatures. However, feeding will most likely not occur in ‘heavy’ seas.

To address these criteria, three preliminary design ideas were generated and proposed during a design review held on 1 September 2004. Those present at the meeting included OOA operational as well as engineering members. One concept was very similar to the one-ton buoy currently in operation. It had a diameter of 15 ft with a large ballast weight tank that was connected to the main structure by four long pipes. Another concept was a catamaran type with two separate hulls. Each hull, holding two feed silos, was connected by a large machinery superstructure. The third concept was a large diameter structure (20 ­ 22 ft) with the ballast weight (concrete) located within the lowest section of the main structure. This was the concept selected and is shown in Figures 15, 16 and 17.

The superstructure contains most of the machinery needed for buoy operation (including: generator(s), batteries, control panel(s) and feed collection hopper). The main structure will hold the four separate feed storage bins and any pumps required for the external feed transfer system.

Instead of using one large rolled cylinder to construct the main structure, an idea to use 12 flat plates to create the structure has been generated (Figure 18). This is the current configuration being used for further study. The difference between the cylinder and flat plate construction may seem enormous, but the difference is minor in relation to the internal systems as well as the hydrostatics of the overall buoy system.

To perform a detailed hydrodynamic analysis the mass (weight) distribution needs to be known. Since not every mass can be determined until the design is complete, only the major sources of mass have been included for the following analyses. These weights are provided in Table 1. The center of gravity (CG) values are given relative to the bottom of the main structure, not the bottom of the ballast structure.

With the weights and CG values, a hydrostatic analysis was performed for the load and light (without consumables) displacements. A heel analysis was also performed to determine the righting arm for the buoy at different angles of heel. The results are given in Table 2 and Figure 19.

Both the metacentric height values (load and light displacement) are positive and over two feet. This is an indication that the buoy is stable. Since all the righting moments are positive, if the buoy heels over, the righting moment will work to push the buoy back to a nominal heel angle. Since there is the possibility that the feed storage bins will be emptied at different rates, a stability calculation was performed for the case that two feed storage bins, on the same side, are empty, while the other two are full. This resulted in a heel of 14°.

To estimate the dynamic response of the design concept, a database of various buoy shapes was used. The database also includes model data as well as full-scale measurements. These shapes can be related to a dimensionless shape factor (SF), which involves displacement and waterplane area. The database includes a range of SF including spar type buoys (SF greater than one) to large diameter can type buoys (SF less than one). The database values can then be Froude-scaled to the same size as the current design concept for analysis.

By calculating the SF, heave and pitch resonant periods can be approximated by interpolating between prior measurements in the database. The shape factor is 0.67 for the current twenty-ton buoy concept. Values interpolated from the scaled database results in heave and pitch resonant periods of 3.89 and 10.95 seconds respectively. In general, 4 second waves typically do not produce extreme amplitudes (unlike storm waves with periods greater then 8 seconds). Therefore, operational activities would be more manageable. In waves with periods longer than 4 seconds, such as storm waves, it is expected that the buoy will be a wave follower.

One of the main challenges with the larger buoy design is the addition of a major system to handle internal feed transfer. Both the existing quarter-ton buoy and one-ton buoy have not required a system of this type because the feed could be stored high in the buoy. Since a twenty-ton feed capacity is required, the feed needs to be located low in the buoy for stability concerns.

Possible types of internal feed transfer systems investigated included: flexible screw augers, straight augers, aero-mechanical conveyors, bucket elevators, cable conveying methods and pneumatic systems. After an extensive review of the options, the search list was narrowed down to flexible augers or pneumatic systems. Flexible augers consist of a flexible plastic pipe with a metal helix (spring shape) inside. The helix is connected to an electric motor that spins the helix in the pipe. Pneumatic systems use moving streams of air and pipe.

Since fish feed pellets have not typically been used in a flexible auger system, a field test was conducted at Flexicon Corporation located in Bethlehem, PA on 9 September 2004. A system comparable with our needs was setup and tested. Two different sizes of feed were tested: 6.5 mm and 13 mm. Both sizes of feed were conveyed without problem at a transfer rate of over 100 ft³ per hr (approximately 3700 lb/hr, depending upon the density of feed being conveyed). This represents twice the design criteria rate. A trip to Eastport, ME was conducted to view a pneumatic conveying system in operation. Pneumatic conveying methods are currently used extensively for fish feed pellet transfer to surface cages as well as for feed transport inside a structure.

A meeting, conducted on 13 September 2004 by OOA engineering and infrastructure personnel was held to discuss the options. Since both systems can fulfill the desired tasks a design matrix was generated to best examine the major differences between the two systems. Using the matrix and discussion, a flexible auger system was decided upon to be the best choice. Some major factors in the decision were cost, system maintenance and complexity. The working concept for the current internal feed transfer concept is shown above in Figure 20, with dimensions shown in Figure 21. The need for four separate storage bins results in four flexible augers.

Shown in Figure 22 is the layout for all four separate flexible auger systems inside the buoy. The augers are curved due to the physical space restraints inside the buoy.

The current internal feed transfer concept starts with feed stored in the storage bins. Each storage bin is connected to the inlet of a flexible auger. The feed is transferred up the auger tube and is discharged to one central collection hopper. Once the feed is in the collection hopper, the external feed transfer system begins. With all four feed storage bins being able to fill the central collection hopper, any storage bin will be able to feed any cage.

Since the UNH aquaculture site uses submerged cages, the feed must be delivered to the cages in a water solution. Due to the size of the proposed buoy design, it will have to be moored separately from the grid, resulting in extremely long hoses (up to 800 feet) running from the buoy to the cages. To best protect the hoses they need to be submerged over the full length, further indicating a need for feed delivered in a water solution.

A major design decision is whether to use a batch or continuous feeding system. Since the design criteria specifies a maximum of 1800 lb/hr (30 lb/min) needs to be transferred, a batch system cannot be used. This is due to the start/stop time (approximately 15 sec) required for the pumps to cycle on and off. With that time the number of batches is limited, making the batches much larger. A larger batch size is not a desirable characteristic. For a continuous system to function, a free surface (water-to-air interface) is needed to accept the supply of feed pellets. Pumping water through a pipe creates back-pressure (head) and causes the water level of free surfaces to rise. The continuous feeding system needs to have a method to control the level of the water to eliminate the possibility of flooding the buoy.

The current concept for a continuous feeding system uses a ‘Mixing Chamber’, two pumps and piping. A schematic of the concept is shown below as Figure 23.

Before any feeding takes place valves (not shown) after pump ‘B’ will open/close to direct the feed into the appropriate discharge pipe to feed the desired cage. Pump ‘A’ turns on to allow water to fill the mixing chamber. Once the water reaches the desired level, any extra water will exit through the two exit (overflow) pipes. These pipes are very large (~6 in) in relation to the inlet pipe (~3 in). The flow rate for the inlet pump is to be twice the exit pump so that the mixing chamber will always have water in it. Once the water reaches the desired level pump ‘B’ will be switched on. Up to this point no feed has been introduced into the system. After the water is moving and the free surfaces that are needed in the mixing chamber are present, the feed can be dropped in. A rotary airlock will control the feed introduction (very similar to the current 1 ton feed buoy).

The three major components of the external feeding system are the two pumps, ‘A’ and ‘B’, and the mixing chamber. Pump ‘A’ is a standard pump that only needs to move water at the desired rate. Pump ‘B’, on the other hand, needs to be a pump that can handle feed passing through the pump with little to no damage to the pellets. Pumps that have been researched include trash centrifugal, centrifugal, fish, diaphragm and positive displacement pumps. First, a trash centrifugal pump was investigated that met the feed transfer rate criteria and was relatively inexpensive.

An experiment was designed to run feed through a trash pump similar to the desired pump to view the effects on the feed. The test was conducted on 16 November 2004 and was witnessed by engineering and infrastructure members. A schematic of the setup is shown on Figure 24.

The results of the experiment were not favorable. The feed had a high breakage rate that was not acceptable. A picture of the conveyed feed is in Figure 25. All of the other types of pumps (positive displacement, diaphragm, centrifugal) except for the fish pump were discarded for various other reasons including: size, cost, power requirements and safety. Fish transfer pumps are designed to transfer live fish without damage. This type is pump is currently being researched.

Since the concept of the “Mixing Chamber” has never been tested, a full-scale test is currently being planned. This is crucial to the project and will be carried out in the near future.

PROGRESS ON TASK #4: DEVELOPMENT OF AN OPERATIONS CENTER
Control/Operations Center Development: Since the last report, the live streaming data from the one-ton buoy has been added the operations website. This has been a great asset allowing detailed troubleshooting without leaving the lab. This data includes; GPS position, all DC battery bank voltages, DC current drain, AC line voltages from the generator, AC line currents from the generator, fuel tank level, feed hopper level, feed hose back pressure, generator oil pressure and water temperature. The live data also allows operators to determine whether any electrical breakers are tripped, which has proven itself very useful when determining corrective actions.

Most of the engineering developments for the control/operations center have been occurring on the offshore side of the system. Significant software development for the buoys to interface with land has occurred over the past year. These include advances to allow the buoys to receive incoming feeder timing and control files automatically on the hour. Previously, an operator would have to break into normal program sequence on the hour and manually upload the control files. A shore based software package for uploading the files to the buoys is almost complete. This development is a large step towards the development of a fully autonomous aquaculture system. After both the buoy and shore software is fully developed and tested a web-based graphical user interface can be developed to allow personnel to monitor and control the operations online, and perhaps allow the public to view certain aspects of the site operation also.

Video Telemetry Development: In early spring, before the quarter-ton buoy was overhauled, the issue of how to obtain quality video telemetry came to the design table once again. This issue has been a difficult one to address since most of the design and construction efforts have been placed on building and maintaining the actual feeder operations. Personnel time has been very limited in the amount that can be spent on video telemetry development. Yet, in early spring a group of engineers got together to discuss the different options available for remote video telemetry. Several methods for transmitting video were discussed in detail including; 900 MHz spread spectrum telemetry, 2.4 GHz spread spectrum telemetry, 900 MHz spread spectrum telemetry upload of high quality recorded video, low speed satellite telemetry, and high speed satellite telemetry. In the end it was decided that 900 MHz spread spectrum telemetry, with possible expansion capabilities to record and upload, would be the most feasible and economical to implement.

During the overhaul of the quarter-ton feed buoy, a 900 MHz spread spectrum radio was installed in place of the prior 2.4 GHz 802.11b wireless system, and the video worked as designed. From the offshore site to the Seacoast Science Center (SSC) a video stream of two to three frames per second was experienced using the system, and when transmitted back to Durham over the SSC’s bandwidth limited T-1 line, the stream was reduced to one or two frames per second. An upgrade of the SSC T-1 line from a 256 kbps line up to a full 1 mbps is suggested to fully view streaming video from the site without bandwidth limitation. The video link worked fairly well until mid-October when the feed hose broke free of the cage, and the video connector attached to the bottom of the hose was torn apart. A quick onsite splice was performed to try to remedy the situation and has worked marginally since then. It is suggested that next summer when the buoy is hauled out of the water for a complete inspection and overhaul that the hose splices be replaced.

In September the one-ton buoy had all of the video wiring installed from the cage up to the buoy and is now ready for the wireless telemetry system to be installed. A new type of radio is being tested for this application. Microwave Data Systems, in Rochester, New York, has developed a 900 MHz spread spectrum transceiver that will allow for up to 512 kbps of data throughput. The Freewave transceivers installed in the _-ton buoy are limited to 115 kbps of throughput, therefore these new radios should almost multiply the bandwidth by 5 times. The system is currently under development, and if all goes well, may be installed for testing before the end of December.

PROGRESS ON TASK #5: FINITE ELEMENT MODELING IMPROVEMENTS
As part of Task #5, improvements to the finite element program are consistently being made. These improvements are described in this section.

Aqua-FE was originally developed to have a commercially available FEA pre-processing/post-processing program Mentat as its graphical user interface. This configuration has been successfully used by Open Ocean Aquaculture Project to analyze various fishcage designs and corresponding mooring systems. However, there are some shortcomings to this arrangement:

• The commercially available program Mentat is manufactured and supported by MSC.Software corporation. It is usually sold in a package with the general-purpose FEA program Marc at a substantial price, and is not available to some users. Thus, for this project’s outreach and dissemination activities there is a need for a stand-alone software program for finite element analysis of fish cages, feed buoys and moorings.

• Program Mentat is periodically updated by the manufacturer; this results in the necessity to introduce corrections to the computer moduli that prepare the input data for Aqua-FE. Also, several steps of data manipulation are required to add the environmental loading information to the finite element mesh data generated by Mentat. A single program developed specifically to serve as the Aqua-FE pre-processing module would make the data preparation more efficient.

To address these concerns, a graphical user interface module Aqua-FEed was developed (Figure 26). It has the capacity to create finite element meshes of truss-, beam- and buoy-type elements, prepare the environmental loading data, and prescribe the computational parameters in the input format of the Aqua-FE simulation software. The module is written in a programming language Python, and is compatible with any PC-based Windows operating system.

Buoy Element Development: In the design of structures submersed in fluids, an often-encountered structural component is an isolated buoyant, or non-buoyant, mass. For example, in the design of offshore aquaculture net-pens, buoys are often utilized in the layout of the mooring system to provide a shock absorbing influence and pre-tension on mooring lines. In addition, objects heavier than water are sometimes hung from net-pen components to provide stabilization. Often, these buoys and weights are spherical or close to spherical. Thus, for finite element analysis, they can be modeled in two different ways. First, a single truss element of suitable length and diameter may be used to provide a roughly similar dynamic response to relative fluid motion as the sphere would exhibit. Second, a separate buoy element may be developed to predict the behavior of a spherical object interacting with the moving fluid. Details of the buoy element developed for Aqua-FE are provided in Kestler (2004).

Feed Buoy Modeling Comparison: The UNH finite element computer program Aqua-FE has been increasingly called upon to simulate the interaction of feed buoys with moored net pen systems. Aqua-FE was used to analyze the design for mooring the _-ton and 1-ton capacity feed buoys, and it will be employed in developing the mooring system for the 20-ton capacity feed buoy. Since the basic finite element utilized by Aqua-FE is a truss element (a bar in tension or compression), strategies have been developed for constructing an array of elements (a “model”) to represent a solid buoy. To evaluate these approaches, Aqua-FE dynamical predictions for an 80-ton capacity feed buoy design were compared with tank testing using a relatively large physical model (the tank testing part has been previously described in the New Feed Buoy: SBIR Phase I section).

The feed buoy design considered was generated in the collaborative SBIR Phase I study with Net Systems and is detailed in the final report by Swanson et al. (2004). The report also summarizes free-release tests in heave and pitch and Response Amplitude Operator (RAO) experiments using a 1:20.77 scale, one m tall physical model having a mass of 21 kg. In similar tank experiments using a much smaller model, described by Ahern (2002), very good correspondence was found between the scale model dynamics and full scale observations. Thus the data set for the larger UNH/Net Systems buoy model is regarded as a very good basis of comparison.

The physical model is shown in Figure 27 (left), and the corresponding Aqua-FE model is illustrated in Figure 27 (right). (Massless truss members needed to maintain truss rigidity are not shown for clarity.) The main cylinder components of the buoy are represented by single elements, while flat planes, such as the ballast and the underside of the main cylinder, are made up of an array of elements. In creating the model, matching mass and volume (buoyancy) was given preference over matching projected area (contributing to drag).

In the free-release tests for heave, the buoy was released upright from a height above its equilibrium position and allowed to undergo damped oscillations. Pitch tests were analogous with the release angle tilted but no change in vertical position. A linear mathematical model was fitted to the time series data to infer virtual mass and mass moment of inertia, damping ratios, as well as damped natural frequencies and periods. Experimental methods, theory and data processing used here followed those described by Fullerton et al. (2004). Inferred dynamical parameters are provided in Table 3 (all results are Froude-scaled to full scale values). Aqua-FE was used to simulate these tests, and corresponding results are also presented for comparison.

The predicted damped natural period for heave, critical for characterizing heave resonance, agrees with tank experiments within 21%. Predicted virtual mass is higher than that measured, while damping rations are lower. Pitch damped natural frequencies agree to within 15%, and damping ratios correspond very well.

In the RAO tests, the buoy was acted upon by a sequence of regular waves at periods bracketing typical wave energy regimes. The buoy was moored using two test configurations designed (for experimental purposed) to minimize the direct influence of the tethers. Figure 28 shows the “no pre-stress” configuration in (a) and the “pre-stress configuration” in (b).

The normalized heave comparisons are shown in Figure 29 and Figure 30 indicate very good agreement in RAO magnitude. The resonant peak, however, is shifted somewhat reflecting the different damped natural periods noted in the free-release tests. The normalized pitch comparisons in Figure 31 and Figure 32, on the other hand, show peaks in the range of 0.05 ­ 0.10 Hz. Maximum pitch RAO’s differ by 32% in the pre-stress case and by less than 6% in the no pre-stress configuration.

The overall conclusion is that Aqua-FE modeling of solid bodies is adequate for the purpose of developing buoy moorings, as long as sufficient factors of safety are employed to compensate for the comparison discrepancies. Thus the important large amplitude capability of the nonlinear Aqua-FE model can be employed seamlessly in mixed systems. The use of truss elements described above reflects the best compromise in matching mass, volume and projected area.

CITED MATERIAL:
Ahern, J., (2002). The Validation of a Wave Measurement Buoy. Master of Science in Ocean Engineering Thesis, University of New Hampshire, Durham, NH.

Baldwin, K., B. Celikkol, R. Steen, D. Michelin, E. Muller, P. Lavoie (2000). Open Aquaculture Engineering: Mooring and Net Pen Deployment. Mar. Tech. Soc. J. Washington D.C. Vol 34, No. 1 pp 53-67.

Baldwin, K.C, Celikkol, B., Chambers, M., Fredriksson, D.W., Irish, J.D., and M.R. Swift. (2003) Open Ocean Aquaculture Engineering. Proceedings of the Oceans 2003 Conference. San Diego, California.

Fredriksson, D.W., M.R. Swift, E. Muller, K. Baldwin and B. Celikkol. (2000). Open Ocean Aquaculture Engineering: System Design and Physical Modeling. Mar. Tech. Soc. J. Washington D.C. Vol 34, No. 1 pp 41-52.

Fredriksson, D.W., M.J. Palczynski, M.R. Swift, J.D. Irish and B. Celikkol. (2003a). Fluid Dynamic Drag of a Central Spar Cage in C.J. Bridger and B.A. Costa-Pierce, editors. Open Ocean Aquaculture: From Research to Commercial Reality. The World Aquaculture Society, Baton Rouge, Louisiana, United States. pp 151-168.

Fredriksson, D.W., M.R. Swift, I. Tsukrov, J.D. Irish and B. Celikkol. (2003b). Fish Cage and Mooring System Dynamics using Physical and Numerical Models with Field Measurements. Aqua Eng. Vol 27, (2), pp. 117-270.

Fredriksson, D.W., J. DeCew, M.R. Swift, I. Tsukrov, M.D. Chambers, and B. Celikkol. (2004). The Design and Analysis of a Four-Cage, Grid Mooring for Open Ocean Aquaculture. Aqua. Eng. Vol 32 (1) pp 77-94.

Fullerton, B. Swift, M.R., Boduch, S., Eroshkin, O. and G. Rice, (2004). Design and Analysis of an Automated Feed Buoy for Submerged Cages. Aqua. Eng. Vol 95 (1) pp. 95-111.

Kestler, K.J. (2004). Development of Special Finite Elements for Dynamic Analysis of Flexible Structures in the Moving Fluid Environment. M.S. Thesis, Mechanical Engineering, University of New Hampshire, Durham, NH 165 p.

McGillicuddy, G., O. Patursson and G. Mulukutla. (2004a). Free Release Experiments on a Physical Model of a Feed Buoy. Report submitted to Dynamics of Moored Systems, UNH Graduate course in Ocean Engineering, M.R. Swift (Instructor), Durham, NH.

McGillicuddy, G., O. Patursson and G. Mulukutla. (2004b). Wave Response Experiments on a Physical Model of a One-Ton Feed Buoy. Report submitted to Dynamics of Moored Systems, UNH Graduate course in Ocean Engineering, M.R. Swift (Instructor), Durham, NH.

Michelin, D. and S. Stott (1996) “Optical Positioning, Instrumentation and Evaluation”, Ocean Projects Course Final Report, Tech 797, Sea Grant, Kingman Farm, University of New Hampshire, Durham, NH, 85 p.

Palczynski, M.J. (2000). Fish Cage Physical Modeling. M.S. Thesis, Ocean Engineering, University of New Hampshire, Durham, NH, 111 p.

Rice, G.A., M.D. Chambers, M. Stommel, O. Eroshkin. (2003). The Design, Construction and Testing of the University of New Hampshire Feed Buoy in C.J. Bridger and B.A. Costa-Pierce, editors. Open Ocean Aquaculture: From Research to Commercial Reality. The World Aquaculture Society, Baton Rouge, Louisiana, United States p. 197-203.

Swanson, K., G. Loverich, L. Gace, M. Smith, M. Chambers, D.W. Fredriksson, M.R. Swift, B. Fullerton, S. Boduch (2004) “Offshore Spar Based Fish Feeder”, SBIR Phase I Final Report submitted to Sea Grant, Department of Commerce under Contract No. DG133R-03-CN-0047.

Tsukrov, I., Ozbay, M., Fredriksson, D.W., Swift, M.R., Baldwin, K., and Celikkol B., 2000. Open Ocean Aquaculture Engineering: Numerical Modeling. Mar. Tech. Soc. J. 34 (1), 29-40.

Tsukrov, I., Eroshkin, O., Fredriksson, D. W., Swift, M.R. and Celikkol, B., (2003) “Finite Element Modeling of Net Panels Using Consistent Net Element”. Ocean Eng. 30: pp. 251 ­ 270.

C. Important Results or Findings
Results are incorporated above.

D. Difficulties Encountered
No major difficulties encountered, other challenges are described above.

E. Anticipated Success in Meeting Project Objectives on Schedule
See Section II below.

F. Reports, manuscripts, and presentations resulting from the project
Reports and manuscripts are included in the body of the report as reference material. The engineering team has also been involved in several other scientific related activities including seminars, presentations at conferences and other open ocean aquaculture proposal activities. These other initiatives are stated below.

Seminars, Presentations and other Journal activities

World Aquaculture Society Conference 2004: A Study of the Structural Integrity of Salmon Net Pen Gear for Deployment in Exposed Locations. March 2004 in Honolulu, Hawai’i, USA. Authors: David Fredriksson, Jud DeCew, James Irish and Igor Tsukrov.

World Aquaculture Society Conference 2004: Automated Feed Systems for Submerged Open Ocean Aquaculture. March 2004 in Honolulu, Hawai’i, USA. Authors: Glen Rice, Rob Swift, Kurt Swanson, Oleg Eroshkin and Stan Boduch Robinson.

World Aquaculture Society Conference 2004: A Submerged, Four-Cage Grid Mooring for Open Ocean Aquaculture in the Gulf of Maine. March 2004 in Honolulu, Hawai’i, USA. Authors: Judson DeCew, David Fredriksson, Rob Swift and Barbaros Celikkol.

Aquaculture Association of Canada Conference 2004: The Evaluation of Inshore Salmon Farm Mooring Technology for Use in More Exposed Locations. October, 2004 in Quebec City, Canada. Authors: David Fredriksson*, Judson DeCew, James Irish, Igor Tsukrov and Vijay Panchang

Aquaculture Association of Canada Conference 2004: The Evaluation of Inshore Salmon Farm Mooring Technology for Use in More Exposed Locations. October, 2004 in Quebec City, Canada. Authors: Judson DeCew, David Fredriksson and Igor Tsukrov

United States Naval Academy: David Fredriksson presented an Open Ocean Aquaculture Engineering seminar on June 11, 2004 at the United States Naval Academy, Department of Naval Architecture and Ocean Engineering.

Guest Editor - IEEE Journal of Oceanic Engineering: David Fredriksson and Igor Tsukrov are guest editors for a special issue on Open Ocean Aquaculture Engineering scheduled for publication in January 2005. Below are the contributions from the University of New Hampshire.

DeCew, J., D.W. Fredriksson, L. Bougrov, M.R. Swift, O. Eroshkin and B. Celikkol. (2005). Numerical and Physical Modeling of a Modified Gravity Type Cage and M