Prediction of mass and volumetric flows in a full-scale industrial waste treatment plant

In this paper a set of mathematical tools are developed and assembled together to assess and predict mass and volumetric flows in waste treatment systems (WTS). The proposed approach is constructed upon a set of data reconciliation methods, influent fractionation routines and process simulations models (and model interfaces) to balance, analyze, reproduce and forecast the behavior of different compounds within treatment facilities. The proposed approach is tested on full-scale data collected after two five-week measuring campaigns at the largest industrial WTS in Northern Europe. Results show that the proposed reconciliation methodology based on the definition of the identity matrix, data curation, the estimation of missing fluxes and the calculation of Lagrange multipliers allowed to close flow and COD, N, P, S and multiple metals (Na, K, Ca, Mg and Al) mass balances. Plant-wide overall mass balance reveals that that: 1) 32% of the incoming COD is recovered mostly as methane, 2) 23% and 65% of COD and N are biologically removed and leave the system via the gas phase, 3) 32 % and 33% of COD and N are stored in the activated sludge, 4) > 70% of P, Ca, Mg and Al are accumulated in the bio-solids stream as precipitates and 5) > 70% of Na, K and S remain soluble and leave the plant via effluent. Plant measurements were allocated into the model states describing soluble/particulate and organic/inorganic loads arriving to the WTS under study. The average deviation between computer simulations and the result of the measuring campaign is 10.5%. This study also shows that the proposed approach is capable to reproduce main streams neutralization, volatile fatty acid production, particulate removal and nitrate denitrification in the anaerobic water line (buffer tank, primary clarifier, pre-acidification tank). It also correctly predicts organics transformation into biogas in the anaerobic granular sludge reactor. Lastly, it is possible to describe biological and chemical N and P removal processes in the activated sludge and the quality of bio-solids after inactivation/dewatering (reject water /cake). A scenario analysis is included showing the potential use of the presented tools under dynamic conditions. This is the first study where both tracking and prediction of multiple compounds in large industrial sites has been done at this level of detail. The tool resulting from this study can serve as starting point for a variety of applications, for instance: holistic evaluation of retrofitting scenarios, advanced plant-wide control strategie and environmental assessment.