(612b) Numerical Study of Sulfur Trioxide Decomposition in Bayonet Type Heat Exchanger and Chemical Decomposer | AIChE

(612b) Numerical Study of Sulfur Trioxide Decomposition in Bayonet Type Heat Exchanger and Chemical Decomposer

Authors 

Nagarajan, V. - Presenter, University of Nevada, Las Vegas (UNLV)
Ponyavin, V. - Presenter, University of Nevada, Las Vegas (UNLV)
Chen, Y. - Presenter, University of Nevada at Las Vegas
Vernon, M. E. - Presenter, Sandia National Laboratories
Pickard, P. - Presenter, Sandia National Laboratories
Hechanova, A. E. - Presenter, University of Nevada at Las Vegas


The present study is concerned with the use of bayonet type heat exchanger as silicon carbide integrated decomposer (SID) which produces sulfuric acid decomposition product - sulfur dioxide. Thermochemical cycles produce hydrogen through a series of chemical reactions resulting in the production of hydrogen and oxygen from water at much lower temperatures than direct thermal decomposition. Energy is supplied as heat in the temperature range necessary to drive the endothermic reactions, generally 750°C to 1000°C or higher. All process chemicals in the system are fully recycled. In the cycle heat energy enters to the thermochemical process through several high temperature chemical reactions. Some amount of the heat is rejected via exothermic low temperature reaction. The whole cycle includes the three following reactions:

I2 (l) + SO2 (g) + 2H2O = 2HI (aq) + H2SO4 (aq) (1)

2HI (g) = H2 (g) + I2 (g) (2)

H2SO4 (g) = H2O (g) + SO2 (g) + ½O2 (g) (3)

One of the most critical processes in the cycle is the decomposition of sulfuric acid decomposition process via very high temperature and highly corrosive environment. Many researches on bayonet heat exchangers have been done in the past. However there are no reports about application of the bayonet heat exchanger as a decomposer for the catalytic sulfuric acid decomposition. The present study considers a bayonet type high temperature heat exchanger and decomposer made of silicon carbide and quartz for the catalytic decomposition of sulfuric acid in packed bed using porous media approach.

Silicon carbide integrated decomposer also called as bayonet heat exchanger was designed by Sandia National Lab (SNL). The integrated acid decomposer combines the function of boiler, superheater, decomposer and recuperater in a single silicon carbide unit. In this scheme teflon is used for the low temperature region and silicon carbide is used for high temperature regions. The advantage of the silicon carbide decomposer over the others is the presence of the recuperator which heats the incoming acid gases. This would minimize the total input energy required to the system. 35 to 40 mol % concentrated sulfuric acid is coming from a sulfuric acid concentrator and then pumped into the inlet of bayonet heat exchanger.

The bayonet heat exchanger can handle both high temperature and low temperature regions in a single unit. The inlet is made of teflon and maintained at temperatures below 200ºC. The water vapor and the sulfuric acid enter the boiler which would heat the sulfuric acid to 450ºC to produce a sulfuric acid vapor. The superheater would heat the sulfuric acid vapor from 450ºC to 700ºC and the decomposer would heat the vapors to the maximum operating temperature plus provide the heat necessary to dissociate the sulfur trioxide to sulfur dioxide and oxygen.

The decomposer region is located in the top of the bayonet heat exchanger. It houses the pellets. The pellets are made up silicon carbide which contain approximately 1wt.% of platinum. The pellets are spherical shaped and their packing is simple cubical packing. The diameter of the pellets is 5 mm.

The boundary condition for the outer wall is obtained from the thermocouples. The upper wall is maintained under adiabatic conditions. The value for density of sulfur trioxide mix with water is obtained from commercial software Fluent 6.2.16. The operation is carried out under atmospheric pressure. Pressure outlet is selected as outflow boundary condition.

The calculated geometry is meshed using the mesh generator GAMBIT 2.2.30. The used mesh is quadrilateral mesh. A mesh independent study was done for different nodes and from the results the proper number of cells and nodes (33036 cells, 35374 nodes) are selected for the future meshing. The fluid flow within the investigated geometry is assumed to be laminar because the Reynolds numbers are less than 700.

The governing equations are solved in the Cartesian coordinate system with a control volume finite difference method. A nonstaggered grid storage scheme is adapted to define the discrete control volumes. In this scheme, the same control volume is employed for the integration of all conservation equations, and all variables were stored at the control volume's cell center. The numerical scheme used in this study is a power-law differencing scheme, and the solver used is a segregated solver. The SIMPLE algorithm is used to resolve the coupling between pressure and velocity. The governing equations, which were discrete and nonlinear, are linearized using an implicit technique with respect to set of dependent variables. The algebraic equations are solved iteratively using an additive correction multigrid method with a Gauss-Seidel relaxation procedure.

The flow calculations of decomposition of sulfur trioxide to sulfur dioxide in the presence of catalyst using porous media approach are performed. The pressure drop obtained in the pebble bed region is about 3000 Pa. The temperature in the region increases from the inlet and goes on increasing in the pebble bed region. As the decomposition takes place the mole fraction of SO3 is reduced in the pebble bed region. The mole fractions of SO2 and O2 increases in the decomposition region as the decomposition take place. According to the results the decomposition percentage of sulfur trioxide (SO3) is found to be 19.11%. The percentage is smaller than that with the assumption of constant outer wall temperature (33%). It is because the average temperature in the pebble bed region is smaller than for assumption of constant outer wall temperature. The results are compared with the experimental results obtained in SNL.