Fig. 3.  Predicted velocity field (left), temperature profile (right) of the reference case.

Fig. 4.  Predicted concentration profile for C4H10  (left) and C4H2O3  (right) of the reference case.

304 Y.  Dong et  al.  / Chemical Engineering  Science 142  (2016) 299–309

Position 1

Position 2

Position 3

Dimensionless pellet coordinate  

Position 4

Position 5

Fig. 5.  Concentration profiles  of  C4H10  (left  top),  C4H2O3  (right  top),  CO2  (left  bottom)  and  temperature  profile  (right  bottom)  inside  the  pellet  at  five  sampling  positions  in the reactor. Position 1: r ¼ 0 m, z ¼ 0.1 m; Position 2: r ¼ 0 m, z ¼ 0.25 m; Position 3: r ¼ 0 m, z ¼ 0.4 m; Position 4: r ¼ 5:25 · 10 — 3  m, z ¼ 0.1 m; Position 5: r ¼ 10:5 · 10— 3  m, z ¼ 0.1 m.

the progress of reaction, the concentration gradient is flattened as expected. The flattened profiles along the reactor are a sign for the decreasing reaction rates along the bed due to lower concentration of the reactants and decreasing temperature along the   bed.

If the entire reactor for n-butane oxidation is packed with the same catalyst material, one can image that the catalysts in the lower part of the bed are not used as efficiently as at the bed inlet (Trifirò and Grasselli, 2014). From an economic point of view, a more structured catalyst packing where catalysts with different kinetics are packed at different sections of the reactor may be a good approach (Guettel and Turek, 2010). Positions 1, 4, and 5 are taken from the same axial coordinate but different radial coordi- nate.  One   can   observe   rather  different   concentration profiles

inside the pellet. Position 1 is almost in the hot-spot region as  one

can see in the temperature plot and that may explain the higher concentration gradient with higher reaction rates. Figs. 4 and 5 show that the catalyst pellet located in different sections of the reactor may experience different conditions (concentration of the reactants and temperature). It is possible that these catalysts are chemically different from each other especially on the surface in

Fig. 6. Effect of the macro-pore porosity εM on the simulated fixed-bed reactor performance  for  n-butane oxidation.

outlet of the fixed-bed. These were calculated as integral average at the outlet:

 N — N

accordance  to the exposed  conditions.  In future work, incorpora-

tion of the surface dynamics in the model can be an insightful refinement of the model.

3.2. Effect of pore structure parameters

In total, there are four pore structure parameters in the model of Wakao and Smith (1962) which can be varied or optimized. In the presented study, only three parameters were studied since the total porosity of the catalyst pellet was kept fixed to 0.5. By fixing the total porosity, the pellet density and bed density were kept unchanged with varying distribution of the macro and micro pores porosity in the pellets. The following question is addressed in this work: which pore structure of the catalyst pellet is more efficient for a given mass of active catalyst and fixed pellet shape. All simulations were carried out at the same operational conditions as for the reference case.

Firstly,  the  macro-pore  porosity  εM    was  varied  from  0      to