Aquaculture Food and Marketing

Economic development in West Virginia is critical, particularly in rural communities where traditional economic activities (principally coal and timber) have declined. Rural economic development has been a focus of the Aquaculture Food and Marketing Development Project (AFMDP). Strategies where aquaculture can impact economic development in West Virginia and adjacent states are: (1) development of mine water resources for commercial production of fish to eat, and (2) use of farm raised fish in recreation. Benefits derived from the work proposed in this document have potential for farmers growing fish for both food and recreation. As such the work will encourage development of a dual market for growers in the region strengthening both components of the industry. Work presented in this proposal is designed to complement the work which has been completed and presently underway.

1. Mine Water Aquaculture Productivity through Year-Round Operation with Recreationally Relevant Species Investigate the productivity of hybrid striped bass, largemouth bass, hybrid bluegill and a heat tolerant strain of rainbow trout in treated mine water to assess feasibility of year-round use of treated mine water.

2. Development of Cost Competetive HFRP Fish Tanks. To significantly reduce costs of HFRP for raceway systems by introducing recycled FRP products as partial substitution of chop-strand mats currently being used for the manufacture of tanks installed in WV. The goal is to combine the easy-assembly design being developed in Grant 6 with lower recycled material costs (Grant 8) to produce HFRP tanks at comparable prices with concrete.

3. Software tool for simulation of a flowing water raceway system. Develop and verify a software tool designed to simulate potential performance of a flowing water raceway system.

5. Protein and Lipid Recovery from Fish Processing By-products The objective is to scale up the protein and lipid recovery batch process to continuous semi-industrial application.

6. Technology Transfer. Information developed from this research and previous research will be integrated into an effort to educate people regarding how aquaculture is conducted and its role in economic development.

Much of the research supported by the Aquaculture Food and Marketing Development Project has involved production and processing of trout. A long term objective is to develop a modular raceway system and improve raceway design for commercial production of food fish in flowing water systems utilizing impaired water.

Development of Honeycomb Fiber Reinforced Polymer Tanks for Field Assembly from Flat Panels .

Several designs to join two flat panels were proposed and studied and the model considered to be best suitable for the current application was adopted for the new connector design. This design utilized a 4x4x4 angle plate with ¼” thickness and 3 bolts, 2 with Φ ¼”, normal head and one with Φ ¾”, sunken head, to join the side panel to the bottom panel. The ¾” bolt also has a threaded hole so that one of the ¼” bolts can be inserted into it. Holes are drilled into the flat panels corresponding to the location of the bolts and the angle plate is placed on the bottom panel. The ¾” bolt is inserted through the angle plate into the bottom panel with the threaded hole in the longitudinal direction of the core. The side panel is placed at one end and is bolted to the angle plate. The other ¼” bolt is then inserted through the side panel into the bottom panel and is tightened into the threaded hole in the ¾” bolt, to form a watertight connection. An elastomeric pad is used between the side panel and the bottom panel to provide an even surface of contact

Based on formulations proposed by Davalos and Chen (2005), the coefficient of elastic restraint was calculated and finite element models were created. Two sets of prototype connections were assembled, one with a 2” bottom panel and the other with a 4” bottom panel. On assembly of the first set of samples, it was decided that modifications to improve stiffness should be implemented before testing the connection. Experimental testing for stiffness and strength of the connection were performed and modified finite element models were created and are discussed in the later sections.

Modified design. The sunken head bolt was replaced with a regular bolt and the thickness of the angle plate was increased from 1/8” to ¼” to increase stiffness. Also, two 3/8” bolts were used to attach the angle plate to the side panel instead of one ¼” bolt. The ¼” bolt which is inserted into the ¾” bolt was also replaced with a 3/8” bolt of the same length. An adhesive sealant was used between the angle plate and the FRP panels to provide perfect contact between the two. A 1/8” steel plate was used as a washer on the outside of the side panel for the 3/8” bolts. Using these components, we assembled two prototype connectors, one with a 2” thick bottom panel and the other with a 4” thick bottom panel.

Experimental Testing of Connection Samples. The two connections assembled were tested first in the linear range and then to failure, to determine the rotational stiffness and strength of the connection. The experimental test setup was similar to the setup used to test the original stiffened and un-stiffened connection samples. The bottom panel is clamped rigidly to a vertical steel column. The side panel rests on a 4” wide support and has an overhanging length of 42”. As in the previous tests, the side panel is loaded at a distance of 36” from the joint and deflections at 24” and 36” are noted. Five strain gages are bonded to the side panel, to record the strain distribution near the joint.

Finite Element Modeling of the Connection. Solid elements are used for the FE analysis of the connection samples for ease of modeling. The side and bottom HFRP panels are modeled to the required dimensions using equivalent properties. A connection assembly is separately modeled as a single entity as shown in Figure 2.

The mesh that is generated for all the components and elements corresponding to the holes on the FRP panels are then deleted. Contact surfaces are created where surface interaction is expected, such as simulating contact between the side and bottom panels. It is also assumed that the sample is resting over 1” thick plate, as observed in the lab, which allows for angular rotation of the joint between the angle plate and the panels. Nuts of the corresponding bolts are tied to the nodes on top of them, simulating the actual experimental condition, where the nuts are in compression and hence can be assumed to be fixed to the panels. Boundary and loading conditions are applied to the model similar to the experimental setup.

Conclusion. By observing the failure of the samples on both experimental testing and finite element modeling, it can be seen that local failure of the material occurs around the washer of the ¾” bolt, for both the two panel thicknesses used. This may be avoided by using a larger washer, which distributes the stress over a larger area, thus increasing the strength of the connection. It can also be seen that the angle plate separates from the bottom panel as the load increases, which permits the washer to cut into the material. This separation of the angle plate from the panels allows for rotation of the connection (θ). Using a bigger washer would solve this problem to some extent, by resisting the separation forces. Thus from the experimental testing and finite element modeling, it can be observed that the 4” thick bottom panel is stiffer and stronger than the 2” thick bottom panel and hence is more suitable for the modified connection. The new connection design developed in this project offers the potential for economic field assembly of raceway systems using flat panels and relatively simple connections designs.

Production evaluation of two commercial diets for rainbow trout in treated mine water. An experiment comparing two commercially available diets will be completed in May, 2005. The control diet is a commonly used fish meal based diet and the second diet is a a commercially available diet made without fish meal. The diets will be assessed in terms of the following: (1) productivity related factors such as feed conversion ratio, growth rate, and survival, etc., (2) quantity and characteristics of solid and dissolved waste production, and (3) concentrations of organochlorine compounds (e.g., polychlorinated biphenyls, dioxins, etc.) in fish flesh grown in WVU’s mine water raceway.

Due to the potential impacts these findings may have on the aquaculture industry in WV, it is important to determine whether a difference in organochlorine body burdens can be detected between fish raised on a fish meal based diet and those grown using an all vegetable diet. In addition to fish production related benefits, vegetable-based feeds are reported to decrease solids and dissolved waste production, as a greater fraction of the total feed is utilized efficiently by the fish. This work is currently ongoing.

Conduct economic analyses to evaluate costs and benefits of using impaired water for the production of fish for food and recreation. As a precursor to determining profitability of mine water aquaculture, a data base consisting of all high-flow mine sites (>1000gpm) was updated. Measurements included measured flows, county location, land-owner, water quality data, treatment on site, power availability, and a production-related risk rating. We have completed an assessment of profitability (from an aquaculture producer’s perspective) and potential state-wide economic development impacts as a result of mine water aquaculture.

Develop designs for increasing the efficiency for removal of solid wastes from the quiescent zone in raceway systems producing trout. Solids removal mechanisms in flow-through aquaculture systems generally consist of quiescent settling in a chamber followed by evacuation through a standpipe, vacuuming, etc. These methods are known to be time and labor intensive and are generally inefficient. Prior to attempting to design and model more efficient quiescent zones and solids removal systems, it is necessary to first establish baseline hydraulic operating characteristics in conventional systems. Acoustic Doppler velocimetry (ADV) was used for the in situ measurement of three-dimensional velocities at discrete locations in flowing fluids, based on the Doppler Principle. In this work, hydrodynamic properties in a typical raceway system and quiescent zone were characterized using ADV technology and conclusions relevant to improving overall operation and future modifications to removal processes in raceways were developed.

Development and verification of a model to describe production of rainbow trout in raceway systems. Work to date on this project has focused on developing Excel-based software to design and/or evaluate an existing raceway. Currently, the software performs calculations for a single tank in the raceway system using user-defined inputs. These inputs include: seasonal water temperature, initial fish size, desired final size of fish, water flow rate, raceway dimensions, nutrient density of the diet, etc. Using accepted growth rate figures and other information (Soderberg 1995, Klontz 1991, and Westers 2003), the growth and optimal feeding histories are calculated along with oxygen usage and nitrogen generation rates within the tank.

Characterize the impact of varying CO₂ and O₂ levels on growth efficiency, nutrient utilization, and fillet attributes of Rainbow Trout and Arctic Charr. Rainbow trout (Oncorhynchus mykiss) grown in water with elevated free CO₂ display lethargic and intermittent feeding behavior, and have slower growth rates compared to fish grown in water with CO₂ concentrations < 25 mg/L. Reasons for the decreased growth are not yet known; however, increased activity levels (i.e during feeding) may result in transient acidosis and reduced oxygen consumption. If the reduced growth of hypercapnic rainbow trout is associated with transient acidosis and reduced respiratory function, then minimizing oxygen demands and activity levels associated with feeding may improve their growth. Feeding smaller meals throughout the day, instead of feeding one large meal in the morning, may help reduce activity levels associated with feeding and subsequent acidosis.

Two strains of rainbow trout (approximately 150 g per fish) were randomly stocked into twenty-four circular, 100-L tanks at a rate of 10 fish per tank. Carbon dioxide and feed frequency treatments were initiated at the beginning of the acclimation period. Fish were acclimated for 14 days to allow for adjustment to the new tank and treatment conditions. Carbon dioxide concentrations and water flows were adjusted as needed to obtain treatment CO₂ concentrations and maintain oxygen levels greater than 80% saturation.

Fish were acclimated to either 20 or 40 mg/L free CO₂ levels. In addition, two control tanks (without fish) were used. Each tank was equipped with transparent, removable polypropylene lids to minimize gas exchange between the water-air interface (Figure 3). The lids have small, snug-fitting holes for passage of the water and gas lines, as well as unsealed hole for feed dispersal and escape of CO₂ gas that might otherwise accumulate. Tanks were fitted with lids and gas lines before stocking the fish. Each CO₂ treatment level was fed one of two feeding frequencies during the acclimation period; either once per day or four times per day. Fish were fed 5% body weight per day (as either one feeding or spread over 4 feedings). Treatment combinations were in triplicate.

On the day of the experiment, CO₂ and feed treatments were continued as during the acclimation period. Feed intake was monitored throughout the 24-h experiment: out-flowing stand-pipes were equipped with netting to catch any uneaten feed. Oxygen and total ammonia-N (TAN) measurements were obtained from the inlet water and out-flowing stand-pipe using handheld meters and standard water quality test kits (YSI, Inc. and Hach Co.). Oxygen and TAN measurements were used for determination of oxygen consumption and nitrogen excretion rates. Carbon dioxide, pH, total alkalinity and total hardness concentrations were measured in the out-flowing stand-pipe using standard test kits (Hach Co.) and standard water quality methods.

Preliminary results in Figure 2 indicate that fish fed once per day (9 AM) had maximum oxygen consumptions earlier (near noon sample) compared to the fish fed throughout the day (9 AM, 11 AM, 1 PM, and 3 PM).

Figure 3. Legend: carbon dioxide level - times fed per day - strain of fish.

Omega-3 Fortified Rainbow Trout The four month feeding trial has been completed at the Reymann Memorial Farm. Fillets recovered from the trout fed a diet with 0, 15, and 23.5% of flaxseed oil were analyzed for moisture, total fat and vitamin E contents as well as fatty acid profile and lipid oxidation. Fillets were also in a storage stability study at 2 and 10°C. The fillets were analyzed for moisture, total fat and vitamin E contents as well as fatty acid profile and lipid oxidation. Preliminary data are presented in figure 4. Omega 3 fatty acid content per gram of flesh on a dry weight basis was 23.9, 17.18, and 46.4 for respective flaxseed oil treatments of 0, 15, and 23.5% at the end of the experiment. The omega 3 to omega 6 ratio was 1.98, 1.81, and 2.41, respectively. This study has been completed and the data analysis is in progress.

Cryprotection and Restructured Fish. Cryoprotectants other than sucrose/sorbitol were evaluated to reduce the sweetness of restructured trout products during frozen storage. Bacterial growth, lipid oxidation, thaw loss, cook yield, color, and texture were evaluated after 1 d, and 3 and 6 m of storage at -20 °C. Sucrose/sorbitol, trehalose, and trehalose/sorbitol, at 8% equally exhibited a cryoprotective action, minimized thaw loss and texture changes, while sodium lactate did not at 2% during 6-m frozen storage. Raw, carbohydrate-treated products had less L* values (Lightness as measured with a Minolta chromameter ) than the control and sodium lactate ones. Following cooking, no difference in L* value was observed. Cryoprotectants and frozen storage time did not affect bacterial growth and lipid oxidation of raw products.

Use of Near Infrared Spectroscopy to predict, real-time, texture development in trout products The purpose of this study was to determine the relationship between NIR spectra and storage modulus (G’, the elastic component) of low-fat beef or trout batters, with or without NaCl. Skinned rainbow trout fillets or unpeeled beef knuckles (IMPS #167) were minced and formulated to contain 10% fat, 30% added water, and either 0 or 2% NaCl. Batters were stuffed into the molds and cooked to 72 °C in water bath. Gels were removed when the internal temperature reached 50, 60, and 72 °C. The NIR spectra between 800 to 1700 nm were collected on these gels. The dynamic rheological property of the batter was performed on a rheometer at 50-Pa stress and 0.1-Hz frequency. The batter was heated from 10 to 72 °C at a temperature ramp of 1 °C/min, and the rheological property was expressed in terms of the storage modulus. Partial least squares (PLS) analysis was used to predict the storage modulus from NIR spectra of the gels from 96 data sets. A PLS cross-validation model used log (1/R) and log (G’). The experiment was conducted with 8 replications. The optimal model conditions for PLS regression occurred with 6 PLS factors at SEC = 0.56 and r = 0.89 (n = 64). When the PLS model was tested against the validation subset, similar performance was obtained at SEP = 0.52 and r = 0.91 (n = 32). During thermal processing, NIR can be used to predict the storage modulus of low-fat meat batters prepared from muscle of two species and at differing salt levels. Application of NIR spectroscopy to texture assessment during thermal processing will facilitate designing muscle foods for consumers with varying needs.

Vitamin E Stability During Fillet Storage and Handling. Fillets were processed from trout fed a diet containing either 200 (a commercial diet) or 5000 (a vitamin E supplemented diet) mg a-tocopheryl acetate/kg for 0, 4, and 9 wk. These fillets were evaluated fresh and after 6 m of frozen storage. Frozen fillets were thawed and stored 3 d at 1 °C prior to analyses. Muscle a-tocopherol of fish fed a vitamin E supplemented diet continuously increased through 9 wk of feeding. Reduced muscle a-tocopherol and moisture, and increased muscle redness and fat were observed in frozen-refrigerated fillets compared to fresh fillets. Thiobarbituric acid-reactive substances were lower in frozen-refrigerated fillets produced from fish fed elevated dietary vitamin E. Proportion of unsaturated fatty acids and omega-3 fatty acids increased as feeding duration increased from 0 to 9 wk.

Supplementing a-tocopheryl acetate (300 and 5000 mg/kg diet) in a trout finishing diet was done to minimize lipid oxidation in oven-cooked fillets and hot-smoked products. Smoking did not affect a-tocopherol content of smoked products compared to raw fillets. Feeding trout diets containing 5000 mg/kg vitamin E increased muscle a-tocopherol content that, in turn, minimized lipid oxidation in oven-cooked fillets produced from fresh and 7-d refrigerated fillets, and in smoked products following refrigerated storage for 8 wk. Dietary vitamin E did not affect fatty acid composition of products from either cooking methods. Oven-cooked fillets produced from 7-d refrigerated, raw fillets and refrigerated, smoked products had lower percents of omega-6 fatty acids and lower omega-3 fatty acids:omega-6 fatty acids ratios compared to fresh, raw samples.

Development of Value-added Food Based on Proteins and Lipids Recovered from Trout Processing By-Products. Based on the initial protein solubility and recovery studies using boneless skinless trout fillets, the trout filleting by-products (frames and heads) were used to recover muscle proteins. A batch system to recover proteins was designed. The recovery system employed a 4.5 L vessel connected with a pH meter and homogenizer for protein solubilization and precipitation steps. The separation was conducted using a batch centrifuge, 6 tubes x 250 ml each tube in one batch. The by-products were homogenized followed by solubilization at pH 2.0, 2.5, 3.0, 12.0, 12.5, or 13.0. Centrifugation at 10,000 x g for 10 min was applied to separate “real by-products” (i.e., bones, skin, fin, scale, insoluble proteins) and fish lipids from the protein solution. The protein solution was recovered and precipitated at pH 5.5 followed by centrifugation at 10,000 x g for 10 min in order to separate the precipitated proteins from the water. Temperature was controlled below 5°C during processing. Ash content in the boneless skinless trout fillets, recovered proteins and the “real by-products” was determined as an indicator of impurity content. The ash content in the recovered proteins and boneless skinless fillets was the same and much lower than in the “real by-products”, indicating good separation of the recovered proteins from the “real by-products”.

The pH of recovered proteins was adjusted to 7.0 and they were used in a silent chopper to develop laboratory protein gels that would allow texture and color evaluation as quality indicators of the recovered proteins. The final moisture content of the gels was adjusted to 80%. The following additives were used during gel formulation: (1) sodium chloride (NaCl) at 2% (w/w), (2) beef plasma protein (BPP) at 1% (w/w) as a protease inhibitor, (3) potato starch (PS) at 3% (w/w) to enhance gel strength, (4) tripolyphosphate (TPP) at 0.3% (w/w) to enhance water retention of the proteins, (5) transglutaminase (T-Gase) at 1% (w/w) to induce protein gelation at low temperature, and (6) sucrose at 4% (w/w) and sorbitol at 4% (w/w) as cryoprotectants to prevent protein denaturation during frozen storage. This formulation resulted in a protein paste that was cooked at 90°C for 15 min. Texture profile analysis (TPA) and torsion test indicated that the gels were firm and cohesive. The tristimulus color values (L*a*b*) indicated that the gels had lower whiteness when compared to the gels made from Alaska pollack grade A surimi. The dynamic test showed that the protein gelation started at 60-65°C, indicating that functional proteins were recovered from the processing by-products.

Protein and Lipid Recovery from Fish Processing By-products: Process Scale-up from Laboratory-batch to Continuous Semi-industrial Application. Based on the initial protein solubility and recovery studies using boneless skinless trout fillets as well as the tests with the processing by-products in the batch system, a continuous semi-industrial system was designed. The system included a continuous homogenizer, two bio-reactors, and two continuous centrifuges. The tests were conducted using trout filleting by-products obtained from our industry collaborator, High Appalachian, LLC (Sophia, WV). Two tests with our industry collaborator, Stephan Machinery (Columbus, OH) were conducted using a continuous pilot scale homogenizer MCH-10. The homogenizer was capable of reducing the particle size to below 0.2 mm and maintaining the flow rate of 120 L/hr, which is suitable for consecutive steps in our process. One test with our industry collaborator, New Brunswick Scientific (Edison, NJ) was conducted using a continuous bio-reactor BioFlo 110 equipped with a 10 L vessel. The bio-reactor was capable of continuous and automatic pH adjustment as well as maintaining the flow rate of 120 L/hr, which is suitable for our process. One test with our industry collaborator, New Brunswick Scientific (Edison, NJ) was conducted using a continuous centrifuge CEPA Z-41. This centrifuge was incapable of separating the protein solution and lipids from the “real by-products” or the precipitated proteins from the water. Therefore, this centrifuge is not suitable for our process. One test with our industry collaborator, Alfa Laval (Richmond, VA) was conducted using a continuous centrifuge GyroTester. This centrifuge was capable of separating our streams. However, the ash content in the recovered proteins was equal to the ash of the trout processing by-products, indicating poor separation. Therefore, more tests have been scheduled for first quarter of 2005 with Alfa Laval using a decanter centrifuge MRNX 438 DD and a clarifying centrifuge. We are currently conducting tests with our industry collaborator, Kendro Laboratory Products (Newton, CT) using a continuous centrifuge CARR Powerfuge Pilot Separation System equipped with vacuum/supernatant pump and vacuum centrate receiver vessel. Another set of separation tests has been scheduled with Westfalia Separator (Northvale, NJ) for the first quarter of 2005.

A second focus of the Aquaculture Food and Marketing Development Project has been to investigate the use of farm raised fish for recreation.

Assess demand and develop marketing strategies for recreational fee fishing packages as complementary recreational activities. Draft of a study which surveyed fishermen who are residents of West Virginia with fishing licenses and fishermen that are not West Virginia residents but have a West Virginia fishing license is nearing completion. The purpose of the study was to assess the level of interest both groups might have in fee-fishing compared to non-fee fishing as a recreational activity. In addition, the study assessed the interest of both respondents in participating in a recreational travel package that included fee fishing, lodging, meals, etc. The results support an interest in both West Virginia fee fishing as a recreational activity and in a recreational fee-fishing package that included lodging, meals, etc. if it were made available.

Data are currently being analyzed for a study that was undertaken to access the level of interest visitors to West Virginia had in recreational fee fishing and a recreational package that included fee-fishing along with lodging, meal, etc. In this study fishing is considered not only as a secondary recreational activity but as a subset or component of a total recreational travel and tourism package for persons interested in vacationing in West Virginia. Preliminary findings show that there is an interest on the part of visitors to West Virginia in a travel and tourism recreational package that includes fee-fishing as one component.

Production of all-female triploid brook trout. Approximately 15,000 all female triploid and diploid brook trout fingerlings were produced at North Carolina State University and transported to the aquaculture facility at Reymann Memorial Farm in May of 2004. In the 11 months since the experiment (two treatments, 4 replicates/treatment) began, there has been no significant difference in growth, survival or feed conversion between the all female triploid and diploid treatments. It is expected that differences among treatments will become most evident during the spawning season in the fall of 2005. These trials will provide a range of baseline production data describing the relative advantages of monosex female and/or triploid brook trout production.

Catch-Related Standards of Quality and Demand Behavior for Fee Fishing Stocking Projects. Hybrid striped bass (HSB) were stocked to compliment the hybrid bluegills that were stocked previously at the same study sites. HSB fishing opportunities were evaluated as part of different recreation program formats (i.e., competitive youth fishing derby, competitive senior’s fishing event, competitive catfish tournament, and drop-in day use). On-site observations and questionnaires were used to evaluate the success of this stocking project at three different WV sites. Among the most desired services reported by anglers include ponds stocked with a variety of fish and at high density levels. HSB were hard hitting and aggressive when first stocked, but later became “hook shy.” The average respondent caught 1 HSB per hour. Over 80 percent of participants were willing to pay $10 per child. HSB were the most likely fish to swallow or partially swallow a hook (16.4% of the time).

Developing and Marketing Fishing Based Travel Packages. This work has only recently begun. The purpose of this study is to develop recreational travel packages that include fee-fishing for visitors who view fishing as a primary motive or as a secondary motive in their decision to visit West Virginia. The research is identifying existing and potential recreational activities, including fee-fishing. The study will also identify lodging and restaurant facilities interested in participating in a travel and tourism recreational package. From the data collected recreational packages for market testing will be developed for specific target markets.

Production of Four Fish Species through Year-Round Operation of a Flowing Water System Utilizing Treated Mine Water.

Four different species have been selected for evaluation in the raceway system utilizing treated mine water at Dogwood Lake. They are: Hybrid striped bass (white bass Morone chrysops x striped bass M. saxatilis) Hybrid bluegill sunfish (Lepomis macrochirus x L. Cyanellus) (male bluegill sunfish x female green sunfish, Largemouth bass (Macropterus salmoides), and Rainbow Trout (Case Western strain). Target harvest density for each raceway unit is approximately 60 kg/m³. Below this density, oxygen concentrations are expected to remain above 70% of saturation and are not expected to impact growth of fish at any point in the system. The number and size of fish in each treatment is described in Table 1. These fish are utilized both for food and recreation and represent the most likely species produced by West Virginia growers for the fishing packages under development. By investigating the growth of these species in treated mine water, opportunities to operate on a year-round basis will be explored and the application of mine waters to support recreational fishing markets will be expanded.

	Hybrid Bluegill (HYBG)	Hybrid Striped Bass (HSB)	Largemouth Bass (LMB)	Rainbow Trout (RBT)
Target Harvest Size (each fish)	1/3 lb	1.5 lb	1 lb	1 lb
Fingerling size (inches)	3to 4	6 to 8	4 to 6	4 to 5
Stocking Number	3000/tank	700/tank	1000/tank	1000/tank

Fish will be purchased from commercial vendors and stocked into the raceway system in September 2006 and grown out for one year. They will be fed a 42% protein, 16 % fat commercially diet by hand with the balance of the ration distributed by pendulum activated demand feeders. The amount fed will be based on the volume of feed distributed to each tank. Feed density (g/L) will be measured and used to convert the volume fed into the weight fed to each raceway unit. Fish will be fed daily to satiation unless water temperatures require less frequent feeding. Fish will not be fed on Sundays or the day prior to sampling. Mortalities will be removed each time the fish are fed and recorded.

Fish will be sampled every six weeks. Total weight will be determined by weighing all fish in each raceway unit. To obtain an average weight, four subsamples will be taken and counted as fish are captured with dip nets and weighed during the process of determining total weight. A subsample of at least 50 fish will be taken to determine condition factor. The length of each fish will be measured to the nearest millimeter, and the average weight of the subsample will be determined. These data will provide direct measurement of feed conversion, average weight, condition factor, and estimation of various parameters during each 6 week period. Fish samples will be taken before stocking, at routine intervals during grow out, and at harvest to assess the potential for bioaccumulation of metals in fish fillets. These assays will be performed according to USEPA guidelines (USEPA 1993).

Production parameters will be compared for each heat tolerant species and with historical data. For instance, historically, for Rainbow Trout reared in treated mine water, the stocking density was 26.4 kg/m³ with an initial loading rate of 0.192 kg/L/min. At harvest, the total net production was 3,275 kg (7,220 lb) with an average loading rate at harvest of 0.40 kg/L/m and a harvest density of 50.2 kg/m³. The survival rate was 98.6%; however, over the duration of the study, 16.6% of the total number of fish stocked were removed for related scientific studies, mortality, and theft. The calculated feed conversion rate (FCR) was 1.4 and the average absolute growth rate over the entire study was 1.52 g/day. The quantitative methods used to calculate productivity-related parameters are:

where M = total fish mass of each segment and V = volume of rearing space of segment.

where M = total fish mass of each segment and Q = water flow rate in each segment.

A schematic of the raceway system is presented in Figure 5. The system consists of two parallel channels on four levels, for a total of eight discrete raceway segments. Each of the eight segments has the following dimensions: 9.1 m (l) x 0.9 m (w) x 1.1 m (d) (30 ft x 3.0 ft x 3.5 ft). The water depth in each segment is controlled by adding or removing dam boards, each of which is approximately 30.5 cm (12 inches) in height. Water is run at a constant depth of 0.9 m (3 ft) at a fixed flow rate of 15.8 L/s (250 gal/min), as determined by rectangular weir measurements. The head loss from the top of the dam boards to the surface of the water in the next segment is 1.1 m (42 inches).

Comparisons of water quality and solids production will be made between heat tolerant species as well as with historic data on rainbow trout collected in previous work. In particular, the following water quality characteristics will be monitored: water and air temperature; water flow rate; five-day biochemical oxygen demand (BOD₅); total suspended solids (TSS); ammonia nitrogen (NH₄⁺-N and NH₃-N); settlable solids; dissolved oxygen (DO); and pH. These parameters are current NPDES regulated water constituents for fish hatcheries in West Virginia. Influent, process water, and effluent turbidity measurements will be taken during field water sampling events. Additionally, alkalinity, acidity, dissolved metals and sulfates will be measured, as each is an important constituent of water chemistry, which can affect fish growth and production. These analyses will add to water quality and fish production data developed over the past three years at the mine water treatment site and will contribute to the modeling project (objective 3). In order to obtain a complete characterization of water quality in the pilot-scale aquaculture system under production conditions, the headbox and final quiescent zones will be outfitted with an in situ water quality monitor capable of measuring pH, temperature, turbidity, conductivity, DO, and depth.

Water samples will be collected at the inlet to the raceway system, the effluent from each quiescent zone and the outlet of each raceway. Emphasis will be placed on sampling of solids from the quiescent zone of the raceways to benchmark differences in solid waste production between individual species. Each parameter will be measured according to the applicable “Standard Method” (APHA 1998) or US EPA (1998) approved analytic method.

The technical challenges of the proposed innovation can be effectively addressed by organizing the research plan into three sequentially major tasks: (1) recycling of FRP scrap for efficient manufacturing of all-recycled and combined recycled and chop strand mat hybrid structural laminates of up to ¼” in thickness, with equal or better properties to laminates made of commercial chop strand mats; (2) manufacturing process development for assembly of core components, consisting of corrugated and straight laminates, and attachment of core to facesheet, to produce prototype sandwich samples with equal or better structural performance than existing designs; and (3) economic assessment of the innovation to demonstrate commercial advantage and market potential. A plan for each of these major tasks with corresponding subtasks is described in this section.

Task 1: Recycling of FRP and Manufacturing and Evaluation of Recycled and Hybrid Laminates: This task will address the following three sub-tasks: (1.1) recycling of FRP materials, (1.2) manufacturing process of laminates, and (1.3) structural evaluations of laminates.

Sub-task 1.1: Production of Recycled FRP Materials - [Performed by KSCI – as Cost-share to this Project]: FRP scrap is the starting point, but it must be processed into materials that are reproducible and which retain fibers long enough to provide the desired physical properties. KSCI will work with Seagull of Volusia County Inc., Edgewater, FL, to produce controlled glass-fibers with statistically defined amounts of either polyester or vinyl ester cured polymers to be used in this study.

Sub-task 1.2: Manufacturing Process of All-recycled and Hybrid Laminates - [Performed by KSCI – as Cost-share to this Project]: The current sinusoidal core configuration and layers of the facesheet for the HFRP sandwich panel are being produced using chop-strand mats (ChSM), for which the WVU team has already established material and system properties. Thus, in this sub-task we will produce flat laminates comparable to ChSM, consisting of all-recycled FRP, if possible, and also recycled FRP and ChSM hybrid combinations. The selected manufacturing process that will be tried is a modified spray-up and contact lay-up combination, based on an existing similar commercial technique (Agarwal and Broutman 1990). This is a low-volume and labor-intense process, but it was selected because of simplicity and manufacturing advantages, including: minimum equipment investment, minimum cost and start-up lead time, low tooling cost, easily implemented by semiskilled labor, flexibility of design, and ability to produce bi-layer laminates (e.g., recycled FRP over ChSM). Since we are interested in reducing material costs, the investigation will include not only glass fibers, but both polyester and vinyl ester resins, suitable for fish tank production.

The feasibility of this manufacturing process will be investigated to produce flat laminates first (Sub-task 1.3), followed by the application of the process to produce and assembled corrugated and flat components of the core, as well as attachment of the core to the facesheet (Sub-task 2.1). KSCI will produce the necessary samples for evaluations by the WVU team.

Sub-task 1.3: Structural Evaluations of Laminates – [Performed by WVU]: The WVU researchers have already obtained material properties for ChSM laminates, which are currently being used for the components of the HFRP panel. This subtask, therefore, will evaluate both stiffness and strength properties of all-recycled and hybrid laminates with three volume percentages of recycled FRP. We describe, materials, testing methods and results.

Materials: The standard core produced by KSCI for fish tanks is made from two 3.0 oz of ChSM which when saturated with polymer cures to a 0.12” cell wall thickness. Similarly, the facesheet consists of three 3.0 oz ChSM layers (0.18”). However, for purposes of material testing, the thickness of the laminate will have to be at least ¼”. Thus, the all-recycled and hybrid laminates will be produced in ¼” thickness, which is equivalent to the volume of a laminate using 12 oz ChSM (about 0.25”). This will allow us to manufacture symmetric hybrid laminates using three volume percentages of recycled FRP, respectively: 25% within two 4.5 oz ChSM, 50% within two 3.0 oz ChSM, and 75% within two 1.5 oz ChSM. It is important to evaluate the best way of manufacturing laminates with recycled FRP, and the material properties that can be achieved; thus, the above four combinations of materials for testing will give us the opportunity to evaluate the concept, and to subsequently select the best process to produce the core and facesheet for the sandwich panel.

Testing and Results: The testing protocol for coupon samples to obtain stiffness and strength will consist of tension, bending, compression, and shear, following modified ASTM guidelines supplemented by proven methods. The number of specimens for each test is estimated as: (6 replications) x (4 sample types) = 24. For tension, samples of 10”x1” will be tested to failure (ASTM D 638-99) to record longitudinal and transverse strains and load-displacement to failure. For bending, samples of 15”x2” will be tested to failure under 3-point loading (ASTM D 790-99), to record mid-span strains and deflections as function of load, following previous work by the researchers (Lopez-Anido, Davalos and Barbero 1995). For compression, the extensive work accomplished and testing tool developed at WVU will be used (Barbero et al. 1999; Makkapati 1994); the samples will be 2”x1”, with strain gages bonded to opposite sides to attain alignment and eliminate bending. Finally, for shear, the Iosipescu test (ASTM D 379-88) method will be used with notched butterfly specimens instrumented with strain gages; the testing tool and the equipment for precise cutting and polishing of specimens are available at WVU. Photographs of ChSM samples tested at WVU are shown in Figs. 2 through 4 for tension, compression and shear, respectively.

Guided by the experiments, the analytical predictions will be based on existing models successfully developed by the researchers. The prediction models for stiffness and strength will be calibrated using the experimental data, which will permit us to make predictions for other possible hybrid laminates to recommend optimum percentages of material combinations, and also to guide the manufacturing of the actual core and attachment to the facesheet.

Task 2: Manufacturing and Evaluation of Prototype Sandwich Samples: This task is organized into 2 sub-tasks: (2.1) manufacturing of core and attachment to facesheet, and (2.2) structural evaluations of prototype sandwich samples.

Sub-task 2.1: Manufacturing Process for Core and Core-to-Facesheet Assembly – [Performed by KSCI as Cost-share to this Project]: Unlike the production of laminates (Sub-task 1.2), the manufacturing of the core is more challenging and will require trials to arrive at a functional solution. The core consists of sequential corrugated and flat components bonded currently by pressure co-curing at contact sections. Both components are produced from ChSM and polymer resin, with a finished thickness of about 0.12”. The core must resist primarily out-of-plane compression and shear, and research at WVU has shown that the thickness of the flat component must be at least about 0.12” (corresponding to 6.0 oz of ChSM) to avoid premature buckling failure. More importantly, the strength of the core-facesheet interface is crucial for product performance; research has shown that a 6.0 oz of ChSM over the inner face of the facesheet can provide an optimum “bonding layer” for embedment of the core into the facesheet by contact-molding. Thus, based on the results of the manufacturing and testing of laminates in Task 1, combined recycled FRP and ChSM hybrids will be manufactured by the spray-up / contact lay-up process (Sub-task 1.2), using the least possible amount of ChSM. The possible manufacturing process for core assembly and core-to-facesheet bonding are described next.

For core assembly, a corrugated component, as thin as possible, will be produced using ChSM (say, 1.5 oz = 0.03”) and existing technology, and while partially cured, recycled FRP will be sprayed over to achieve a prescribed thickness of the hybrid laminate. Similarly, a thin and partially cured flat component will be sprayed with recycled FRP to a given thickness. Then, the smooth ChSM side of the hybrid corrugated component will be co-cured by embedment pressure over the face of the flat containing the recycled FRP. Then, a mirror image of a hybrid corrugated component will be co-cured to the smooth side of the same flat component, with the contact face of the corrugated component containing the recycled FRP (figure 9). This process will be continued to assemble the core for a given product.

For core-to-facesheet assembly, a thin ChSM layer will be applied over the inner face of the facesheet laminate, and then, recycled FRP will be sprayed over to a given thickness, to subsequently embed the core into this “bonding layer.” The rotational and peeling restrain that can be developed at the interface are important to increase core buckling capacity and interface fracture toughness to prevent delamination.

The proposed manufacturing processes will require trials and preliminary evaluations. One concern, for example, is the ability to produce actual-size material (not coupons) of consistent thickness with minimum voids and relatively smooth finished surface; the thickness discrepancies at the contact bonding surfaces between corrugated and flat components can lead to misalignments and dimensional problems, exacerbating core buckling (figure 10). However, the manufacturing of the facesheet using combined recycled FRP and ChSM bi-layer materials should be easier to accomplish.

Sub-task 2.2: Structural Evaluations of Prototype Sandwich Samples – [Performed by WVU]: The WVU researchers have conducted extensive studies of the current product to evaluate its performance to failure, based on experimental and analytical approaches. Based on the most effective manufacturing process defined in Sub-task 2.1, prototype sandwich samples will be produced for testing under compression, flat-wise tension, and bending, with materials and testing methods and results as described next.

Materials: The specimens for compression and flat-wise tension will consist of single-cell units, cut from a sandwich panel. This in-plane symmetric unit of 4”x4” (see Fig. 6) is considered a “representative volume element.” For bending, the sample will consist of a 4” wide strip extending 28” along the sine-wave of the corrugated component, resulting in seven contiguous unit cells. The thickness of all the samples will be 2”, in order to correlate results with existing information obtained for similar samples, and the lay-up of the facesheet will consist of combined ChSM and attached inner “bonding-layer” containing the recycled FRP.

Testing and Results: The testing protocol for strength evaluations will consist of compression, flat-wise tension, and bending, following modified ASTM guidelines and previous methods developed by the WVU team. The number of specimens for each test is estimated as: (6 replications) x (2 sample types, possibly) = 12. The bending failure load is sensitive to the interface bonding layer effect and the contact thickness of the core components, which maybe variables that will need to be valuated experimentally, and predicted by an existing failure criteria using FE modeling.

Task 3: Economic Assessment and Commercial Potential: This third and final task of the Research Plan will address economic aspects for product commercialization. The aquaculture industry will benefit from the present proposed study to substantially reduce costs, while promoting recycling technology of discarded FRP.

Based on the current KSCI core geometry, we anticipate that the core wall thickness, containing about 50% recycled FRP and polymer resin by volume, will have to increase by some percentage to compensate for the expected lower compression and shear strengths of the hybrid laminate. Thus, assuming all currently used materials are priced equally, and further assuming that the value of recycled FRP will be $0.20/lb or $400/ton including transportation and processing, the price of the hybrid laminate, in relation to a currently used 4.5 oz of ChSM priced at $1.19/board-foot, would be about $0.46/bf, representing a substantial savings of $0.73/bf or 61%.

Task 1 – Continuing Development of Software Tool

It is proposed that the software tool continue to be developed in an Excelä programming environment using the VBA programming language. This software platform offers an extensive range of built in functions and user-interface tools and produces a user-friendly interface.

The next phase of the modeling is to expand the analysis to multi-tank raceways. The approach will be to develop software that gives the user as much flexibility as possible in configuring the raceway. The logic flow diagram for the proposed software is illustrated in Figure 11. This figure shows how a user may specify any raceway system consisting of any number of tanks in series with multiple parallel raceways. The user will also provide information pertaining to the location of the raceway, e.g., elevation, seasonal temperatures, water quality, etc. It is envisioned that the software will give the user the capability to simulate scenarios in which fish cohorts may be placed in different tanks at different times. Therefore, the user may input information at any time during a scenario regarding the movement of fish from one tank to another, and the partial or total removal of fish from a tank. The program will then simulate the growth, optimal feed rate, oxygen consumption, nitrogen production, oxygen replenishment via weirs and/or other reoxygenation technologies, and economic parameters such as the cost of feed and fingerlings and the revenue from fish sales.

In order to allow the user to obtain growing scenarios that are as realistic as possible, the software will be written as an event-based tool. Therefore, the user (fish grower) will be able to enter information regarding fish movements within the raceway at any time during the simulation. As much flexibility as possible will be built into the software to enable the user to go back and change the time at which a fish moving event (purchase of new cohort, movement of cohort from one tank to another, or the removal and sale of a whole or part of a cohort) occurs. Optimal feeding rates will be determined based on water temperature. The growth of the different cohorts will be predicted by the software, and basic economic information on the cost of feed, cost of fingerlings, and revenue from sales as a function of time will be included in the output.

Figure 11: Schematic diagram of information flow in the multi-tank raceway simulation software

Task 2 – Analysis, interpretation, and synthesis of data from on-going research

In developing the single-tank model, in precious work (FY 2004), certain gaps in information regarding the growth of trout have become apparent. For example, Soderberg (1995) discusses the growth of a variety of fish in terms of the water temperature and the corresponding feeding rate required to obtain optimal growth. This information forms the basis of growth prediction in the software. Hardy and Barrows (2001) indicate that growth may be tailored by changing the feed rate from the optimal level. In order to implement this additional level of sophist