Design of a New Generation of Rearing Pond Providing for Rapid Removal of Fish Waste

Final Report

Pisciculture Gilbert

AIMAP 2011-Q07

Abstract

Pisciculture Gilbert Inc has built the second generation of self-cleaning rearing pond (Gilbert type) based on the results of numerical simulations performed with FLUENT flow-modeling simulation software. The first generation of this pond was built at the fish hatchery (step 1). Its efficiency in recovering the total phosphorus excreted by the fish was 30%, based on 2010 technology. This design was reproduced as rigorously as possible in the software (step 2), leading to a theoretical phosphorus recovery of 37.4%. This efficiency has therefore become the reference value against which the various second-generation designs are compared (step 3). The optimal modeled design in this second generation of ponds comprises three 2 m by 2 m settling cones, a 2.5% slope in the rearing zone (24.4 by 6.1 m), and a 0.3 m baffle under the upstream recirculation pipe. It yields a modeled phosphorus-recovery efficiency of 43.8% to 45.1%, depending on particle-size distribution, at a water inlet rate of 0.2 m/s, corresponding to two "Fresh-flo", and a 50/50 distribution of recirculation flow upstream and down in the settling zone.

A similar design with a 1.85% slope was built at Pisciculture Gilbert Inc. in fall 2011 (step 4). In actual use with farmed fish at the hatchery, the total maximum average phosphorus recovery achieved was 37.9%. Monitoring pond water quality also made it possible to validate that this new pond model is suitable for raising fish for the stocking market. Based on a comparative assessment of the development and operating costs of two other types of rearing ponds, the additional costs for the new model of the Gilbert-type pond will be recouped in less than 5 years.

1. Introduction

The emergence of new environmental targets in Quebec, such as the Stratégie de développement durable de l'aquaculture en eau douce au Québec (STRADDAQ), whose objective is to reduce phosphorus waste by 4.2 kg per ton of fish produced by 2014, has caused fish-hatchery production in Quebec to drop. Consequently, pond designs applying effective strategies for managing and recovering phosphorus at low cost to fish farmers (Morin, 2004) must be put forward. Therefore, a design must be developed that optimizes water quality while improving environmental effectiveness, which means a unit that can remove or recover manure on an ongoing basis.

In attempting to achieve this goal in the first part of the project, modeling software was used to simulate the efficiency of a new generation of rearing pond by modifying the design parameters. This new model of rearing pond, intended for farming fish for the stocking market, combines water recirculation with the rapid recovery of manure and, inherently, of particulate phosphorus as well as increases the environmental efficiency of fish hatcheries.

During the second part of the project, the new rearing pond was built according to the optimal design parameters arrived at through modeling. The environmental effectiveness of this new pond was then measured under actual production conditions. These data also made it possible to refine the modeling data.

2. Methodology

Overall understanding of the project can be made clearer by defining five major steps that led to the design of the new generation of rearing pond.

• Step 1: The first generation of the Gilbert-type pond was based on Marcotte's work (2010) and tested at the Gilbert fish hatchery.
• Step 2: In 2010, École Polytechnique used this first-generation data to define the variables for the modeling program (SORDAC project; approval pending).
• Step 3: In 2011, École Polytechnique used the modeling program to define and mathematically validate the new parameters for the second-generation of the Gilbert-type pond.
• Step 4: In fall 2011, the second-generation of the Gilbert-type pond was constructed and tested under actual hatchery conditions.
• Step 5: Early in 2012, the modeling program was again used to validate field trials based on the test results from step 4.

Various measurements (water velocity, effectiveness measurements) were taken on the pond built in 2010 during step 1 in order to calibrate the modeling software (step 2). Once the pond had been modeled, FLUENT was used to simulate phosphorus recovery. Obviously, the efficiency of the modeled recovery cannot be identical to that of an actual pond, since certain operating parameters—such as the influence of the fish and particle solubility—are difficult to model.

The budget must be calculated to determine the efficiency of phosphorus recovery. Indeed, it has been estimated that 70% of the total phosphorus excreted by fish is particulate in form and that the particle size varies. Consequently, the distribution is real and therefore predetermined. Three budgets were therefore used for the modeling in step 3 based on the theoretical size distribution. The particle-size distributions were based on data from the literature as well as qualitative analysis of feces sampled in 2011. Sindilariu (2009) stated that 70% of the particles excreted by trout are larger than 63 µm. Pfeiffer (2008) obtained a distribution of 26.3% of particles in the 105 to 250 µm range and 28.3% in the 250 to 500 µm range. Moreover, he demonstrated that 17.2% of the particles were larger than 500 µm, that 28.3% were smaller than 100 µm, and that only 1% was smaller than 23 µm. The particle sizes in the modeling in step 3 were limited to four different groups based on three budgets in order to effectively simulate recovery with different behaviors or meal types. Indeed, certain meals result in more cohesive and therefore larger feces than other meals.

Equations are used to determine the recovery efficiency of the total phosphorus excreted by the fish. Estimating that 100% of the particles are trapped (reach the settling cones) yields a recovery efficiency of 70%.

In step 3, in an effort to design a pond yielding maximum phosphorus recovery, several models were tested to determine the optimal design parameters. Simulation variables included the number of recovery cones, the slope of the bottom of the rearing zone, pond hydraulic conditions (such as baffle use and variations in water velocity in the rearing zone and above the cones).

In step 4, the new pond was built at the Gilbert fish hatchery according to the optimal parameters are determined in step 3. Measurements were then taken on the pond in order to determine the phosphorus recovery efficiency (balance of phosphorus production and recovery) and to determine if the rearing conditions (water quality) were adequate for fish production. Lastly, a comparative analysis with two other types of conventional, open-circuit rearing ponds was carried out in order to determine if the new model make good financial sense.

In step 5, in order to refine the mathematical model, a more exhaustive budget was conducted, taking into consideration a new particle-size distribution and the measurements taken in the field during step 4. Indeed, 10 sizes were used in the simulation. Moreover, phosphorus recovery calculations were simulated in order to better understand the impact of potential changes on certain design parameters, namely changes to the bottom slope in the rearing zone and the rate of flow of recirculation water.

3. Analysis and discussion

3.1 Step-3 Simulations

The best results for modeled design 2 (test 7) was obtained when the water inflow velocity at the weir located at the entrance to the rearing zone was 0.4 m/s. This involves using four Fresh-flo-type aerators. Given the space requirements for such aerators, it was decided to use only two. Modeled design 3 made it possible to validate that the additional investment for sloping the bottom of the rearing zone would compensate for using fewer Fresh-flo-type aerators. Sloping the bottom of the rearing zone and using a lower water velocity in the recovery-cone area would make it possible to carry off the larger particles rolling along the bottom and reduce the amount of smaller particles being pulled towards the water outlets.

Modeled design 4 was used to build the pond in step 4 of the project. In addition to yielding the best efficiency for recovering the total phosphorus excreted, it provides good results for most hydraulic combinations. This design is similar to design 3. The difference is the addition of a baffle under the water outlet upstream of the settling cones. This baffle directs a larger proportion of particles to the settling zone. Supposing that half the rate of recirculation flow of 187.2 m³ is directed to the settling zone, the superficial load over the cones is 7.7 m³/h/m². Reducing this load by decreasing the rate of flow of recirculating water, the simulated design provides for recovering more 130 µm particles than 750 µm particles. As a result, the larger particles fall to the bottom too soon and, therefore, are not transported to the settling zone. Modifying the design parameters in modeled designs 5 and 6 did not further increase recovery of the total phosphorus excreted by the fish.

3.2 Field Measurements in Step 4

The measurements taken on the pond built in step 4 were analyzed from three standpoints:

• Summary of phosphorus production and recovery: An average phosphorus recovery efficiency of about 19% was observed in tests 1 to 3. This low value could be accounted for by the meal having a higher phosphorus content. It was considered that the proportion of dissolved phosphorus excreted by the fish after consuming this meal was higher than that resulting from the meals used in the other tests. The recovery efficiency therefore increased from about 29% to 38% during test 4 to 10 with a more conventional meal containing 1% phosphorus. The highest efficiency was obtained during tests 4 to 6 with the addition of a third aerator. In comparison to the simulations, increasing water flow in the pond resulted in higher phosphorus recovery in the cones. This increased velocity appears to push particles that fall to the bottom of the rearing zone towards the cones. During the simulations, these particles were considered lost;
• Quality of rearing water: Based on the data, the water quality remains adequate for fish farming with a fish density of 6.4 kg/m³. Given the winter conditions, it was not possible to increase this density in order to check the water quality with a higher density of fish. Degassing was nevertheless required to eliminate the CO2 from the new water. Specific care should be considered to maintain oxygen at the minimum recommended levels.
• Comparative financial analysis: The development and operating costs of a Gilbert-type pond would be $26,250 higher and $6,000 lower compared to an all-earthen conventional pond and $10,000 higher and $4500 lower than for a conventional concrete-covered pond. In dividing these additional development costs by the savings resulting from operating costs, the analysis shows that the additional costs for setting up a Gilbert-type pond would be recovered in 2.2 to 4.4 years.

3.3 Step-5 Simulations

The optimal design, which was used to build the pond, has been simulated, with the addition of a third aerator during step-4 testing. Unlike in field testing, the addition of an aerator during simulations consistently yielded a significant reduction in the simulated recovery efficiency of particles measuring 300 µm or less. This could be accounted for by the difficulty in perfectly modeling particle flow in the pond. Diversifying particle size, however, shows that the recovery efficiency of particles measuring 300 µm or more is nearly 100%, except when the simulated design used a single aerator and maximum efficiency was obtained. Increasing the slope of the bottom of the rearing zone always resulted in increased phosphorus recovery. Even so, the results of the tests performed on the actual pond were similar to the results simulated with the mathematical model.

4. Conclusion

Using numerical modeling software made it possible to shed light on the effect of combining construction and operating parameters of a rearing pond, thereby making it possible to increase the efficiency of recovering the total phosphorus excreted by fish. Indeed, the second-generation of the Gilbert-type pond was built and tested, yielding a maximum phosphorus recovery efficiency of 37.9%, compared to 30% for the first generation of this type of pond. The second generation has three settling cones, a bottom slope of 1.85% in the rearing zone, three aerators generating a recirculation flow of 270 m³/h, and a flow distribution of 50%/50% between the water outlets upstream and downstream from the settling zone. Field testing demonstrated that this pond provides adequate water quality for fattening fish for the stocking market. When compared with more conventional rearing ponds, the additional development costs for the Gilbert-type pond are recovered in less than five years, while significantly reducing operating costs. Despite several differences, primarily with respect to the number of aerators to use, the modeling software was highly useful in designing the pond.

References

ANSYS FLUENT 12.1 in Workbench User's Guide. 2009.

Marcotte, D., 2010. Nouveau concept d'étang d'élevage avec réutilisation de l'eau et enlèvement régulier des boues. SORDAC, Document de transfert de technologie No. 2010.3.

Morin, R., 2004. La production piscicole au Québec. Document d'information. Ministère de l'Agriculture, des Pêcheries et de l'Alimentation, Québec.

Pfeiffer, T.J., Osborn, A., Davis, 2008. Particle sieve analysis for determining solids removal efficiency of water treatment components in a recirculating aquaculture system. Aquacultural Engineering, Vol. 39, Issue 1, pages 24–29