Demister configurations

The purpose of the Demister configuration -study was to evaluate the effect of different demister support structures on the flow profiles inside the separator vessel, particularly at the demister inlet. Three different support structures were studied, one of which was the reference demister frame of the base geometry. The structures are presented collectively in Fig. 46.

FIGURE 46. Three demister support structures studied in the Demister configuration -study. Refer to Fig. 22 for dimensions. Demister pad illustrated in blue, horizontal plate in

red.

In the Demister configuration -study two different modifications to the base geometry were studied. Modification 2A includes a horizontal plate placed at the same elevation as the upper edge of the demister pad. Modification 2B has the horizontal plate placed at the elevation of the lower edge of the demister pad. The plate is ring-shaped and extends from the demister frame to the vessel wall. In every case, the demister pad is held in place by a 30 mm wide upper support ring and a 50 mm wide lower support ring. The demister frame

extends 50 mm below and above the demister pad in Modifications 2A and 2B, respectively. Two inlet distributors were used in the Demister configuration –study, the Impact Plate Type 1 and the T-Junction. Only high design values were simulated. The maximum and the standard deviation of the vertical velocity at the demister inlet for each support structure and inlet distributor are presented in Figs. 47 and 48, respectively.

FIGURE 47. Maximum values for vertical velocity 5 cm below the demister pad in m/s in two modified Demister configuration -cases with two different inlet distributors with high

design capacity. Base geometry case results presented as reference.

FIGURE 48. Standard deviation of vertical velocity 5 cm below the demister pad in m/s in two modified Demister configuration -cases with two different inlet distributors with high

design capacity. Base geometry case results presented as reference.

0

Modification 2A does not significantly change the values of maximum velocity and standard deviation at the demister inlet. Both indicators obtain slightly inferior values as compared to the Base geometry, with the exception of maximum velocity for the T-Junction which is slightly improved. The effects of Modification 2B are much more dramatic. Flow uniformity as well as the maximum velocity values are clearly improved for both distributors. Most notably, the T-Junction which produces a markedly inferior flow profile with the Base geometry actually produces a more uniform pattern than the Impact Plate Type 1 with the Modification 2B. A summary of vertical velocity planes 5 cm below the demister pad for each Demister configuration case is presented in Fig. 49.

FIGURE 49. Vertical velocity profile 5 cm below the demister inlet in m/s in each modified Demister configuration case. Base geometry results presented as reference, high

design capacity in all cases.

Fig. 49 reveals that Modification 2A amplifies problems around the demister frame already encountered with the Base geometry. This problem is described in Chapter 9.1.6 concerning the T-Junction distributor and is caused by the returning flow traveling downwards near the lip of the demister. As illustrated in Fig. 34, this creates a vacuum zone around the outer edge of the demister that blocks flow through that particular part of the demister. No such problem exists with Modification 2B because the horizontal plate prevents the flow from looping around the demister frame. Since there is no “lip” below the demister pad in the Modification 2B, the effect of the sharp angle of the demister frame is not visible on the plane in Fig. 49. Although the T-Junction produces a very even flow profile 5 cm below the demister inlet, the sharp edge of the demister frame causes local velocities to increase on direct contact with the demister.

The velocity acceleration due to the demister frame and the vertical velocity profiles on a plane in direct contact with the demister pad are presented in Fig. 50. This phenomenon has a strong effect on the demister inlet velocity profiles of the Modification 2B configurations, where the velocity increases right at the entrance to the demister. The effect on the Right Angle distributor is not as strong as on the T-Junction: the velocity increases around the same region, but the local velocities do not exceed 4 m/s. The horizontal planes are best compared with Fig. 49 to see how much the flow profile changes in the last 5 centimeters before entering the demister in each configuration.

FIGURE 50. Acceleration of flow velocity due to demister design in the Demister configuration cases. Left: Horizontal plane, Right: Velocity profile at demister inlet level.

Refer to Fig. 26 for plane locations.

In general, sharp corners should be avoided especially around key areas like demisters since they have the capability to accelerate flow velocities if the flow is forced to turn around the corner. The problem becomes more manageable if the flow is only traveling past the corner without having to change direction.

The Modification 2B seems to improve the flow profile produced by the T-Junction more than by the Impact Plate Type 1, to the point where the flow profile produced by the T-Junction is actually more uniform of the two studied distributors. This advantage is lost if the entrance to the demister is not rounded since the flow created by the T-Junction needs to make a sharp turn to enter the demister. Provided that the flow acceleration problems are addressed either by rounding of the demister entrance or by distributor design, the Modification 2B seems to be the best option among the studied Demister configurations.

Along with leveling the flow profile, it prevents the formation of the no-flow zone along the demister frame. The extent of this benefit is uncertain though, because the demister support rings always restrict the flow along the outer edge inside the demister.

These results show that no distributor can be declared universally inferior to any other design as the overall performance of the separator unit is always dependent on all of the components inside (and often also upstream of) the vessel. Different combinations of structural components can result in flow behavior that is not always readily predicted if only a single component such as an inlet distributor or a demister is considered.