70

40 50

Pressure drop (arbitrary units)

FIGURE 9.9 Comparison of predicted pressure drop with plant data (from Ranade, 1999b).

High

FIGURE 9.10 Flow and mixing in the mixing cup (from Ranade, 1999b). (a) Contours of stream function. (b) Left side: contours of oxygen mass fraction; right side: contours of ethylene mass fraction. (Legend not shown due to confidentiality constraints). Reproduced in colour plate section between pages 210 and 211.

which of course depends on the relative flow rates of air and the mixture of ethylene and HCl. It can be seen that (Fig. 9.10) the fluid escaping through the annular open space travels almost vertically upward. The maximum turbulent kinetic energy is located near the bottom orifice. The region of highest turbulent viscosity (and, therefore, of highest mixing rates) is, however, somewhat away from the bottom orifice (near the top of the two recirculating loops at the bottom of the mixing cup). The maximum value of turbulent viscosity is almost 5000 times the value of molecular viscosity. These detailed flow and composition results were analyzed to identify regions containing mixtures within the flammability envelope and variations with design and operating parameters. The possible accumulation of ethylene and HCl in the re-circulation loops near the bottom of the cup was also examined quantitatively.

The validated model for the mixing cup was then used to examine the possibilities of operation at higher flow rates and with higher oxygen concentrations. At higher oxygen concentrations, the flow rate of the ethylene stream becomes higher than that of the air stream and the stagnation point moves downwards. Two important design concerns (the possibility of an enlarged bottom orifice due to corrosion, and the possibility of a non-symmetric distributor pipe in the mixing cup), were examined using the computational model with both ordinary and oxygen-enriched air. The influence of an asymmetric feed pipe for the ethylene/HCl stream on flow and mixing within the mixing cup was also quantitatively examined. These results were useful in understanding the fluid dynamics and the limits of the existing configuration of mixing element and distributor for the ethylene/HCl stream.

(b) OXY reactor: When simulating overall flow in the OXY reactor, it is computationally intractable to resolve scales of the order of the bottom orifice of the mixing cup. Therefore, the resistance offered by such small openings and associated abrupt direction changes within the mixing cup were represented by a sub-model (based on the concepts of flow through porous media). Simulation results of throughput versus pressure drop across the mixing element were used to adjust the inertial resistance coefficient of the porous media (assuming that the contribution of laminar resistance is negligible). The value of this inertial resistance coefficient naturally depends on the diameter of the bottom orifice of the mixing element (for a corroded bottom, the value will be lower) and other construction details.

Overall flow simulations of the reactor were carried out to simulate normal operation as well as operation with malfunctioning mixing elements. Typical predicted flow results are shown in Fig. 9.11 in the form of contour and iso-surface plots. The flow simulations give an insight into operation of the OXY reactor. Air fed to the reactor first flows in a downward direction. After impinging on the reactor bottom, air is distributed evenly in the bottom portion below the grid and enters the mixing elements. It mixes with ethylene and HCl and escapes from the annular openings (with a velocity of the order of 5 m s-1) of the mixing cup. It can be seen that the jets from the annular opening reach up to the ethylene and HCl stream distributor (see Fig. 9.11b). In order to examine the influence of reactor internals and the layout of mixing elements on gas flow above the grid plate, a volume above the grid plate was modeled separately. It must be noted that flow above the grid plate may be significantly influenced by the presence of solid particles. The model used in the present work can be extended to simulate gas-solid flows using recent advances in the understanding of the kinetic theory of granular flows. However, here, the scope was restricted to understanding the role of layout of mixing elements on the grid in distributing gas in the reactor. The high velocity jets emanating from the mixing elements dominate the immediate region above the grid plate. Typical predicted results of simulations of gas flow above the grid plate using single-phase flow equations are shown in Fig. 9.12. These predicted results were then used to calculate particle trajectories to obtain useful information about particle impingement on reactor internals and possible implications for the erosion of internals. The model and results were used to examine alternative layouts of mixing cups on the grid plate to minimize mal-distribution and re-circulating regions above the grid plate. The results were also used to understand and to evaluate possible de-fluidization of catalyst particles near the wall region.

The model of the overall reactor was used to examine various scenarios, such as a cup with enlarged bottom orifice or the total absence of a cup. This type of malfunctioning mixing element (enlarged orifice or absent mixing cup) offers a point of

High velocity jets from mixing cups

FIGURE 9.11 Flow characteristics of oxyhydrochlorination reactor (from Ranade, 1999b). (a) Contours of axial velocity, (b) iso-surfaces axial velocity (0.75 m s-1), (c) contours of oxygen mass fraction (legend not shown due to confidentiality constraints).

least resistance to the airflow. This leads to substantially higher velocity jets escaping the grid plate (with velocities up to 80 m s-1). Such high velocity jets may lead to significant mal-distribution and erosion within the OXY reactor. Such localized high velocity flow provides significantly more oxygen in that region, which may be of concern from safety and selectivity (reactor performance) point of views. The computational flow model presented here provided much more quantitative information about the gas flow in the OXY reactor. Not all the quantitative results can be discussed here for reasons of brevity and confidentiality. The results presented here may, however, give an essence of what kind of analysis can be carried out using the detailed computational flow model to evaluate the influence of various operating parameters on flow, and therefore on reactor operation. The computational model allowed not only quantitative estimation of the limits of existing reactor hardware, it also allowed evaluation of alternative configurations (mixing element/layout) to improve these limits. Despite certain limitations (since reactions and some other aspects were not included in the model), the computational model was found to be quite helpful in engineering decision making to realize performance enhancement of an existing

High velocity jets from mixing cups

not shown due to confidentiality constraints). Reproduced in colour plate section between pages 210 and 211.

OXY reactor. The results obtained and the flow model form a sound basis for further work on the modeling of gas-solid flows in fluidized reactors. It may be possible to include reactions in such a model to develop a comprehensive tool to enhance OXY reactor technology. The methodology can be used to enhance both existing and new reactor technologies.

0 0

Post a comment