Determinação da não-idealidades de biorreatores tipo lagoa anaeróbia

Abstract

The conventional anaerobic pond-type reactors are widely used for wastewater treatment, just as in urban areas as in rural areas. This is because of the simplicity and low cost of deployment. When dimensioning anaerobic pond, it is necessary to adopt a kinetic model of anaerobic digestion process and a material balance for the reactor. Ideal models, like continuous-stirred tank reactor (CSTR) or Plug Flow Reactor (PFR), are generally used for material balance. However, due to the size of the ponds and the inherent hydrodynamics, they must be provided deviations from ideality. Through the analysis of residence time distribution (DTR), this research aims to determine the non-ideality of the reactor anaerobic pond-type for a typical reactor’s geometry, evaluating the effect about residence time project. Also, it was aimed to propose a model of balance material that incorporates the non-idealities of this system. For this, it was used the technique of pulse injection tracer (methylene blue) that, with the measurement of the concentration of it at the output, the DTR function is determined. From these data sets, it was predicted a model with non-ideality of the reactor. To perform the experimental runs, it was constructed a wood reactor and internally lined with PVC geomembrane of 8 mm, with ratio L/D of 2/1, ratio H/D of 1/4, edge inclination 1:2 (D/4:H) and total useful volume of 0,92 m 3 . The experiments were conducted adopting residence times project of 3, 6 and 9 days. The experimental normalized values obtained for residence time medium, variance and asymmetry, raised the chances that (i) there is dead volume in the reactor; (ii) closest flow of a CSTR in detriment of a PFR; (iii) the flow of this reactor can be represented by means of CSTR reactors in series. Thus, models were theorized, here classified as (i) conventional, (ii) bypass and (iii) recycle. From the analytical solution of each one were obtained the mathematical representations for the DTR that characterize the modeling of non- idealities reactor. These models were adjusted to the experimental data through the nonlinear optimization algorithm called Trust-region-reflective optimization algorithm. As a parameter of comparison, to decide the best model (representation of reactor flow), they were computed the adjusted coefficient correlation. The best adjust in four sets of experimental data was obtained with the model of two CSTRs in series with variable volumes and dead volume. Finally, to the reactor studied, it is proposed a material balance model, able to predict more accurately the conversion of organic matter from wastewater and, when necessary, biogas productivity, supporting the dimensioning of such equipment.