CHEMICAL RECOVERY

Optimum Surface Condenser Design Allows Maximum Evaporator Capacity

A properly designed shell and tube surface condenser will allow a set of black liquor evaporators to operate at maximum capacity with minimal pressure drop on the condensing side. Improperly designed condensers cause loss of capacity and can create serious tube vibration problems leading to tube failure.

This article examines several different condenser designs and the impact of those designs on evaporator capacity and condenser performance. Heat Transfer Research Inc.'s (HTRI) program CST was used to evaluate three different surface condenser designs. The impact of shell-side design changes was determined for a fixed condensing capacity and surface area. The CST program was used to predict each individual condenser's thermal performance and potential for tube vibration problems.

OPTIMAL DESIGN. A condenser is required to condense vapor from the last effect of a black liquor evaporator set. Typically this vapor is approximately 130F (2.19 PSIA). At this relatively low pressure, the vapor specific volume is relatively high. The condenser must condense the vapor and consume as little pressure drop as possible while processing the vapor.

The preferred surface condenser would ideally have the following attributes:

• Minimal pressure drop on the vapor side

   

• Adequate baffles to minimize tube vibration problems

   

• Adequate surface area to condense the required vapor load

   

• Maximum separation of foul components into the "foul" condensate.

Most surface condensers designed today are vertical units with condensate segregation capability built in. The incorporation of condensate segregation in the design adds minimal additional cost to the condenser when it is built. If segregation is not required, the two condensates can be combined outside the condenser and treated as a single stream.

Adding condensate segregation to an existing condenser can be an expensive, difficult, and non-optimal process. It is highly recommended, therefore, that any new condenser be provided with condensate segregation to allow for future changes in how a particular condensate must be treated.

There are a number of different surface condenser designs currently offered by various vendors. The majority of these are predominantly cross flow (X-shells) designs, shown in Figure 1.

The actual design is often a hybrid combining cross flow with areas in the condenser operating as E-shells or J-shells. Most surface condenser vendors offer proprietary condenser designs specifically tailored to the need to segregate condensate and precool the condenser non condensable gases (NCG's) prior to exiting the condenser.

Consequently, many of the condenser designs currently offered commercially cannot be directly modeled on the CST program. However, the three designs evaluated in this article can be used as a guide to indicate the trends to be expected.

DESIGN CASE. Three different styles of shell-X-shell, E-shell, and F-shell designs-were evaluated to determine the impact on evaporator performance. The effect of internal precoolers was not addressed in this analysis.

A hypothetical vapor load was assumed to design each condenser. It was assumed that 100,000 lb/hour of vapor from the last effect of a black liquor evaporator was to be condensed in a shell and tube condenser.

In a properly designed evaporator set, this vapor would be 130F and 2.19 PSIA. Cooling water would be supplied at 85F and would exit the condenser at 115F. The pressure of the vapor exiting the surface condenser was held constant at approximately 2.04 PSIA (127F).

The pressure and temperature of the vapor entering the condenser were adjusted to maintain a constant condenser exit pressure. The surface area of the condenser was held constant for all three cases. The number of tubes, tube length, tube-wall thickness, tube pitch, and tube pattern were all held constant. The type of shell or baffling was changed and the impact on the operation of the condenser and the evaporator set was evaluated. The hypothetical black liquor evaporator set was assumed to operate with a total of 106F of total delta T available. The case by case data are shown in Table 1.

X-SHELL DESIGN ANALYSIS. The pure cross-flow design, or X-shell, is shown in Figure 1. In the cross-flow design the vapor enters the middle of the shell and flows along the length of the shell. Tubes are usually removed from a portion of the shell to allow passage for the vapor to travel along the length of the shell. The vapor then flows across the condenser bundle and is condensed.

In this design the vapor crosses the bundle only once. Additionally, the vapor is distributed across the entire length of the bundle so that no one section of the bundle is forced to handle large quantities of vapor. Tube supports can be added as required to minimize tube vibration with no increase in pressure drop across the bundle.

In the hypothetical surface condenser selected for this exercise this design has a pressure drop of 0.18 PSI. This represents a temperature loss of 3F, or about 3% of the total 106F total delta T available. The actual coefficient for this design was 252 BTU/hour-ft2F vs a required coefficient of 250 BTU/hour-ft2F. This represents a 1% overdesign based on area.

E-SHELL DESIGN ANALYSIS. In the E-shell design (Figure 2) the vapor zig zags back and forth across the bundle as it traverses the length of the shell. The vapor enters the top of the bundle and is condensed as it works its way to the bottom of the shell.

In this design the entire vapor throughput is forced across and through the bundle. The hypothetical E-shell surface condenser had a pressure drop of 1.09 PSI. The inlet pressure to the condenser rose to 3.12 PSIA (143F) vs the desired pressure of 2.19 PSIA. This represents a total loss of 16F or 15% of the delta T available for the entire set.

The actual coefficient for this design was 238 BTU/hour-ft2F vs a required coefficient of 185 BTU/hour-ft2F. This represents a 29% overdesign based on area. In this particular design the condenser appears to be overdesigned by 29%. This is actually caused by the excessive pressure drop inherent in the E-shell design. The unit appears to have too much surface area. In fact, the evaporator set capacity is diminished by 15% because of the pressure drop of forcing the vapor through the condenser.

HTRI defines several conservative guidelines to indicate when a careful consideration of possible tube vibration should be made. Two of these criteria are the following:

• The ratio of the unsupported tube length to the TEMA maximum span is greater than 0.8.

   

• Cross flow velocities (average, bundle, and baffle tip) are greater than 80% of the critical velocity.

The E-shell design used in this analysis assumes baffle spacing at the TEMA maximum. The ratio of the unsupported tube length to the TEMA maximum span is 1.0, which exceeds the conservative HTRI criteria of 0.8 listed in the first item above. The baffle cut was also set at the maximum allowed by HTRI of 49% of the diameter. With this baffle design the baffle tip cross velocity and the average cross-flow velocity are three times the critical velocity in the inlet zone of the condenser.

The best way to reduce the cross-flow velocities would be to increase the baffle spacing, but this is already at a maximum. The baffle cut is also at a maximum and cannot be further modified. This E-shell condenser shows a high probability of tube vibration problems with a design that is already at maximum baffle spacing and baffle cuts. In essence, there is no baffling arrangement with this E-shell that will not result in a high probability of vibration problems.

F-SHELL DESIGN ANALYSIS. In the F-shell design (Figure 3) the vapor zig zags back and forth across the bundle as it traverses the length of the shell just as it does in the E-shell. The vapor enters the top of the bundle and is condensed as it works its way to the bottom of the shell.

In this design, however, a longitudinal baffle splits the condenser into two passes on the shell side. Vapor enters the top of the shell and flows downward through the first pass. It then passes through the baffle and flows back up the shell for the second pass.

The hypothetical F-shell surface condenser had a pressure drop of 2.26 PSI. The inlet pressure to the condenser rose to 4.30 PSIA (157F) vs the desired pressure of 2.19 PSIA. This represents a total loss of 29F or 27% of the delta T available for the entire set.

The actual coefficient for this design was 262 BTU/hour-ft2F vs a required coefficient of 144 BTU/hour-ft2F. This represents an 82% overdesign based on area. In this particular design the condenser appears to be overdesigned by 82%. Again this is actually caused by the excessive pressure drop inherent in the F-shell design. The unit appears to have too much surface area. In fact, the evaporator set capacity is diminished by 27% because of the pressure drop of forcing the vapor through the condenser.

The F-shell design used in this analysis assumes baffle spacing at the TEMA maximum. The ratio of the unsupported tube length to the TEMA maximum span is 1.0, which exceeds the conservative HTRI criteria of 0.8. The baffle cut was also set at the maximum allowed by HTRI of 30% of the diameter.

With this baffle design the baffle tip cross velocity and the average cross-flow velocity are two times the critical velocity in the inlet zone of the condenser. The best way to reduce the cross-flow velocities would be to increase the baffle spacing, but this is already at a maximum. The baffle cut is also at a maximum and cannot be further modified.

Like the E-shell design, the F-shell shows a high probability of tube vibration problems with a design that is already at maximum baffle spacing and baffle cuts. In essence, there is no baffling arrangement with this design that will not result in a high probability of vibration problems.

The F-shell design is rarely encountered in black liquor evaporators. What is sometimes encountered are two E-shell condensers in series. The F-shell design demonstrates the increased pressure drop and resulting capacity loss of operating E-shells in series.

EXISTING CONDENSER RETROFITS. There are numerous, older condensers in operation today that use the E-shell design. These units can exhibit very high pressure drops, consuming 2.5 in. of mercury pressure drop on average. The pressure drop can be even higher depending on the specific baffling provided in the condenser. A loss of 2.5 in. of mercury represents an 18F loss or 17% of capacity in our hypothetical set of evaporators.

An existing condenser with high-pressure drop can be converted to a hybrid E-shell/X-shell by the addition of vapor inlet nozzles. This is an effective method to gain evaporator capacity without adding equipment. The addition of one or more vapor inlets along the length of the shell redistributes the inlet vapor load along the length of the shell, effectively converting the unit to a hybrid cross-flow design.

In some instances, it is possible to gain several percentage points of evaporator capacity with this design change. There is no equipment cost associated with this retrofit, since no additional surface area is added. Vapor ducting modifications are required, and nozzles must be cut into the shell of the existing condenser.

TIPS AND DESIGN CONSIDERATIONS. The following are several concluding tips for choosing the optimum surface condenser design.

• The preferred condenser design is one that is predominately cross flow in nature. Hybrid X-shell and E-shell designs are acceptable as long as the E-shell component of the design is kept to a minimum.

   

• E-shells and F-shells have considerably higher pressure drops relative to X-shells and consume considerably more of the system delta T.

   

• Avoid operating condensers in series. Certain circumstances and process requirements will dictate series operation, but it should be avoided where possible.

   

• E-shell and F-shell designs have considerably higher probabilities of tube vibration problems relative to X-shells.

   

• Design all new condensers for condensate segregation, even if it is not currently needed.

   

• Consider constructing the internal precooler baffles of stainless steel. These baffles have been known to corrode with time, causing extensive bypassing of hot vapor to the vacuum system.

   

• The lower two to four feet of the shell should be constructed of stainless steel. The condensate/vapor interface in this region can be an area of high corrosion.

   

• If at all possible, locate the bottom tube sheet at a sufficiently high enough elevation to gravity flow condensate into the hotwell. This is normally about 40 feet above grade. This will eliminate the need for a condensate pump.

   

• Consider having the condenser vendor supply platforms that are supported off of the condenser itself. Platform access should be provided to the top water box and the bottom water box as a minimum. Other access may be required depending on the specific design of the condenser.