Estudos sobre a modelagem e simulação de um reator snox

One of the main environmental problems in industrialized countries is related to release into the forming gas environment of acid rain. The amount of greenhouse gases released into the atmosphere is controlled by the environmental legislation, that is becoming increasingly restrictive. Several processes are applied to industry to remove these contaminants from currents resulting from industrial processes. Among these, there is the SNOX® Haldor Topsoe process, which is capable of removing NOx and SOx from industrial emissions, aiming to frame the concentration of these gases within the parameters required by environmental legislation. The removal of these gases is conducted by the process reactor consisting of two catalytic beds distinct in series. The first catalyst bed has a monolithic catalyst structure, in it occurs the reduction reaction of NOx. The second bed has a catalyst with shape of cylindrical rings, in this bed occurs the oxidation of SO2. In this work was modeled and simulated the dynamic behavior of atmospheric emissions abatement reactor in similar operating conditions to a reactor of a SNOX® unit. Each of the catalyst beds, that it make up the reactor, was modeled and validated separately. In the first bed was considered the reactions of reduction of NO and oxidation of SO2, this mathematical model was simulated and evaluated in the operating conditions presented in Tronconni et al. (1998). In the second catalytic bed was considered that only the reaction of oxidation SO2 occurred in the bed, the mathematical model of the oxidation bed was simulated and evaluated in operating conditions presented in Almqvist et al. (2008). Both models showed good fit to the experimental data with error between 2-11%. Thus, the two models had been written in computer code and simulated in reactor operating conditions SNOX® shown on Schoubye and Jensen (2007). The first catalyst bed had an increase in temperature along it of 9 ° C and the conversion of NO was 92%. In the first bed, the inlet region was the region with larger amount of catalytic sites occupied by ammonia. On the second catalytic bed, the SO2 conversion was 96,6%. As part of the modeling and simulation study for reactor the global sensitivity analysis techniques were applied aiming to determine the influence of some parameters over the concentrations of NO, NH3 and SO3 in the bed of catalytic reduction of NO and the concentration SO2 and SO3 in SO2 oxidation bed. From this study it was determined that among the evaluated parameters which it had the greatest influence on the output levels of the first reactor bed was the length of this bed. From the mathematical model developed in this paper we were also generated data to infer the concentrations of NO and SO3 in the reactor outlet, which was performed by neural networks and support vector machines. It was found that the performance of infer carried out by the neural network resembles to the performance of support vector machines.