(484h) Dynamic Modeling of Recirculating Aquaculture Systems with an Integrated Application of Nonlinear Model Predictive Control and Moving Horizon Estimation

World aquaculture fish production has reached a highest record of 82.1 million tonnes in 2018 (FAO, 2020). Ecological concerns involved with this fast-growing food sector such as high water usage and untreated wastewater release can be addressed in recirculating aquaculture systems (RAS) where the addition of a treatment section allows for 90–99% of the water to be recycled and reused (Martins et al., 2010). Despite the low contribution of RASs in total fish production, the application of these environmentally friendly systems to produce fingerling and grow-out species is increasing. For instance, Denmark and Netherlands produce 12000 and 9680 MT/year of African catfish/eel and trout through established RAS technologies respectively (Martins et al., 2010)

Mathematical modeling is an important tool for better understanding, analysis and further success of RAS. There have been several attempts towards model development of these systems in recent years. Wik et al., (2009) proposed an integrated dynamic aquaculture and wastewater treatment framework for RAS where fish growth, gastric evacuation, feed requirement, and biological reactors have been modeled. However, a growth model that takes into account the effect of water quality parameters directly on fish performance were not considered in that study. Karimanzira et al., (2016) presented a model for a RAS coupled with a recirculating hydroponic system for the production of Nile tilapia and tomatoes. The fish response to RAS policies such as degree of water recirculation or the effect of different operating conditions such as temperature or dissolved oxygen were not addressed in that study.

In the present study, a dynamic mechanistic model for recirculating aquaculture systems is introduced to assess the functionality of RAS as a sustainable technology in fish culture. Growth performance of fish is a key element that directly impacts the economic viability of commercial aquaculture facilities (Baer et al., 2011). To account for this condition, our model explicitly considers fish growth and mortality relations to predict the fish well-being under different environmental conditions (e.g., temperature, oxygen concentration) and management strategies (e.g., make-up flow rate to fish biomass ratio, feeding regime). Having access to this information provide insights on the preliminary stages of RAS design and assist farmers and stakeholders in making the best decisions regarding the operation of RAS. Therefore, our model represents dynamic equations for fish (gastric evacuation, growth, mortality) as well as the major units involved in RAS (fish tanks, mechanical filter and biological reactors) to predict the waste generation and performance of the treatment section, and their impacts on fish welfare. The proposed model was validated using experimental data available in the literature and agreed with multiple observations reported in previous studies.

Typically, RASs are often subject to external perturbations and changes in the operating conditions. Hence, the RAS performance in closed-loop is critical for maintaining water quality parameters at acceptable values for fish growth. In addition to the water quality constraints inherent for fish well-being, the nonlinear and interactive behavior of RASs would require the application of advanced model-based control strategies such as nonlinear model predictive control (NMPC). To the authors’ knowledge, no study has reported the application of NMPC to RAS systems. To demonstrate the potential of NMPC for RAS applications, the closed-loop behavior of the system was investigated by making changes in the oxygen set point in the fish tanks and introducing disturbances through fish removal. Under these scenarios, oxygen concentration in the fish tanks and biological reactors as well as the ammonia and nitrate concentration in the fish tanks were selected as the controlled variables. The aeration rate in the fish tanks and biological reactors along with the make-up water flow rate to the fish tanks were considered as the variables available for control. A moving horizon estimation (MHE) was embedded within the closed-loop RAS control scheme to provide the NMPC framework at each sampling interval with adequate estimates to those RAS states that cannot be measured online (e.g. chemical oxygen demand, ammonia concentration). The MHE framework was chosen since it can explicitly account for constraints on the water quality requirements of fish as well as the nonlinearity of the system. The results showed that the dynamic operability of the RAS can be maintained at their set-points in the presence of disturbances. Hence, it is expected that such closed-loop framework would enhance fish performance and the overall RAS profitability.

References

Baer, A., Schulz, C., Traulsen, I., & Krieter, J. (2011). Analysing the growth of turbot (Psetta maxima) in a commercial recirculation system with the use of three different growth models. Aquaculture International, 19(3), 497–511. https://doi.org/10.1007/s10499-010-9365-0

FAO. (2020). The State of World Fisheries and Aquaculture 2020: Sustainability in action. FAO. https://doi.org/10.4060/ca9229en

Karimanzira, D., Keesman, K. J., Kloas, W., Baganz, D., & Rauschenbach, T. (2016). Dynamic modeling of the INAPRO aquaponic system. Aquacultural Engineering, 75, 29–45. https://doi.org/10.1016/j.aquaeng.2016.10.004

Martins, C. I. M., Eding, E. H., Verdegem, M. C. J., Heinsbroek, L. T. N., Schneider, O., Blancheton, J. P., d’Orbcastel, E. R., & Verreth, J. A. J. (2010). New developments in recirculating aquaculture systems in Europe: A perspective on environmental sustainability. Aquacultural Engineering, 43(3), 83–93. https://doi.org/10.1016/j.aquaeng.2010.09.002

Wik, T. E. I., Lindén, B. T., & Wramner, P. I. (2009). Integrated dynamic aquaculture and wastewater treatment modelling for recirculating aquaculture systems. Aquaculture, 287(3), 361–370. https://doi.org/10.1016/j.aquaculture.2008.10.056