Modeling and Simulation of RAS

Other conference contribution, 2011

From a system point of view, recirculating aquaculture systems (RAS) are in general more complex than flow-through systems. Not only because they may include quite extensive water treatment but also because of the feedback interactions between the water treatment system, the feed, the fish and the control systems attached. As a consequence, their behavior can be difficult to predict, analyze and also to control. In particular for RAS in land based fish tanks, where the water exchange should be as small as possible, there is a strong feedback and high demands on the water treatment, e.g. the maintenance of an efficient nitrification, denitrification and organic removal. Modeling and simulation can be an important tool to deal with this increased complexity, and move the development of RAS further towards an ecologically sustainable fish production. Many different models have been developed to describe fish growth, energy needs, gastric evacuation, feed conversion etcetera. However, when RAS are modeled these models are in general connected to simplified and stationary models of the wastewater treatment (WWT). Within the research area of biological wastewater treatment on the other hand, advanced models of different treatment stages have been developed, though without aquaculture in mind. In general, the dynamics is then considered because in most applications the conditions the bacteria are exposed to vary intrinsically depending on, for example, changes in weather conditions, time of the day and day in the week. RAS are also intrinsically varying, and never in a true steady state, simply because the fish is growing and, hence, the load on the treatment is also increasing. In the work presented here we suggest a framework for integrating the two modeling cultures (Aquaculture and WWT). This means that the waste from fish and feed has to be described in terms of the components needed for the wastewater treatment models. It also means that some additional processes and components needed for aquaculture have to be added to the standard WWT models. Some practical experiences from the development of a RAS simulator are addressed and as an illustration a RAS for Eurasian Perch is modeled and simulated.