Introduction
Recirculating Aquaculture systems (RAS) are gaining importance for a sustainable future seafood production. Reasons are the increasing threat to coastal ecosystems and climate change that is imposing thermal distress on organisms in conventional systems.
RAS could be a solution if the water circulation can be largely closed. This requires control over nutrient concentration such as nitrate from ammonia excretion of fishes converted in the biological nitrification. Nitrate accumulating in RAS process water is often controlled by strong water exchange contradicting the sustainable concept of RAS. Without denitrification RAS pollute the environment as open aquaculture installation do.
Heterotrophic denitrification operated with an organic carbon source is an efficient but potentially instable process. The carbon supply to denitrification is crucial to avoid accumulation of nitrite that is an unwanted intermediary product in biological denitrification. By detecting the occurrence of elevated nitrite concentrations in the denitrification effluent the dosage of organic substrate could possibly be adjusted to meet with the demand.
This abstract summarizes company-based research. It describes the discovery of a virtual nitrite sensor that is capable to detect a deficient carbon supply to denitrification. With that a new opportunity is revealed for more sustainable RAS technology.
Material and methods
The investigations were carried out in a precursor model of a SEAWATER Cube. For details refer to Orellana and co-workers (2014, Aquacultural Engineering 58, 20–28). The RAS was fully automated. Process data were saved for later use and evaluation.
The 8 m3 RAS included a subsequent treatment of the denitrification effluent in a floatation to remove excess bacteria biomass, to replenish oxygen, and to oxidise unwanted inorganic and organic residuals (Fig. 1).
The denitrification process was maintained by feeding acetate as carbon source. The acetate dose was estimated from literature data. It was continuously fed into the denitrification by pulses of constant volume (Fig. 1, metering pump).
Nitrate-rich process water was fed into the denitrification whenever the Oxidation-Reduction-Potential (ORP) in the denitrification decreased to values of - 110 mV and below. Water flow stopped at - 90 mV ORP. The effluent water from denitrification was passed into the flotation for clarification and removal of excess bacteria biomass.
The flotation was operated with ozone. The two-point regulator adjusted the ozone feed to maintain an ORP at around + 400mV in the flotation (Fig. 2). In this experimental set up a reductive milieu was followed by an oxidising milieu in the flotation delivering nitrate free and oxygenated process water back into the RAS (Fig.1).
During experiments ORP was continuously monitored. To investigate the effect of ozone, nitrate, nitrite, and carbon feed on the denitrification process and ORP the concentrations were either determined discontinuously (O3, NO3-, NO2-) or calculated from mass flow data (carbon source).
To investigate the effect of elevated nitrite on the ORP within the flotation the filter was disconnected from the RAS and nitrite was dosed into the flotation. Subsequently ORP, ozone, and nitrite concentrations were measured.
Results
Figure 2 shows a short time section of the registered process data revealing two recurring patterns of the ORP signal in the flotation filter. The typical up and down movement of the ORP signal was caused by switching on and off the ozone generator according to the measured ORP within the flotation filter (Fig. 1, A).
Time by time it was observed that the ORP unexpectedly decreased in the flotation filter. This was likely linked to an incomplete denitrification process. Parallel measurements of nitrite concentrations proved that nitrite concentration in the denitrification filter always had increased when a drop in ORP was recorded.
The untypical ORP pattern slowly disappeared if carbon dosing was upregulated (Fig. 2 A, 17:30). The elevated nitrite concentrations disappeared from the denitrification effluent meaning that the full denitrification process had been restored.
Fig. 3 shows the result of an additional experiment to investigate the effect of elevated nitrite concentrations on the ORP in the flotation.
The results (Fig. 3) shows that ORP in the flotation remained on low levels (≦ 300 mV) when nitrite was present in the water. Nitrite concentrations of around 0.6 mg · dm-3 already impeded an increase in ORP. After 18 minutes nitrite had disappeared. Subsequently ORP and ozone concentration abruptly increa-sed at a constant feeding rate.
Discussion
Coincidental observations in a process chain of a denitrification biofilter and a flotation led to the discovery of a virtual sensor that allows to observe the function of a microbial denitrification process. The release of nitrite from denitrification due to an insufficient carbon supply led to a marked change of pattern in the ORP signal of the downstream flotation. The virtual sensing of elevated nitrite marks a next level in the control of denitrification which is inevitable in future RAS operations. No other solution is seen as direct measurements of carbon, nitrate, or nitrite concentrations would be more costly.