Comparação do comportamento hidrodinâmico de reatores anaeróbios de manta de lodo e fluxo ascendente (UASB) em diferentes escalas

Abstract

The aim of this work was evaluate the hydrodynamic behaviour and hydraulic characteristics of upflow anaerobic sludge blanket reactors (UASB) in bench, pilot and full scale. Reactors R1 (1 L), R2 (160 L) and R3 (18,800 L) were operated with hydraulic retention time (HRT) of 6 h, 10 h and 6 h, respectively, treating sanitary sewage. Stimulus-response assays, pulse type, with Eosin Y as tracer were carried out by Passig (2005), Hattori et al. (2011) and Martins (2012) to determine the hydrodynamic behaviour of R3, R1 and R2 reactors, respectively. For the three reactors, the hydraulic characteristics were calculated and the statistical analysis was done with the statistical tests of two-way Anova and Tukey to verify if the scale-effect influences on the hydrodynamic behaviour of the reactors. Results indicated the long tailing phenomena in the curves of concentration varying along the time and an advance of the peak response of the tracer in the residence time distribution curves (RTD) (except in R3). The three reactors studied showed laminar flow regime, with N equal to 3, 2 and 5 to R1, R2 and R3, respectively, for the Continuous Stirred-tank Reactor (N-CSTR) in series model. The average coefficients of Low Dispersion (LD) and High dispersion (HD) were 0.236 and 0.146 to R1, 0.258 and 0.158 to R2 and 0.105 and 0.079 to R3, respectively. Reynolds number verified to R1, R2 and R3 was 0.67, 10.29 and 100.9, respectively, indicating laminar flow regime in the reactors. Volume of dead zones to R1, R2 and R3 was 0.3%, 17.2% and -50%, respectively. It was noted the presence of short-circuiting in R1 (assay 2) and R2 (assays 1 and 2). Among the three reactors studied, it was observed that R3 presented good hydraulic efficiency ( = 8.1). In the remaining assays, the behaviour ranged from satisfactory (0.5 > < 0.75) for assays 1 and 3 of R1 and unsatisfactory ( ≤ 0.5) for the remaining assays. ANOVA and Tukey tests showed statistical difference among the three-uniparametric models (N-CSTR, LD and HD), indicating that the scale influences the hydrodynamic behaviour of the reactors, strongly in R2, followed by R1 and R3.