Depending on the size of the tube bundle, different kinds of tubes are then selected accordingly. This idea results in a lower cost by replacing the enhanced tubes in the top section with the less expensive plain tubes. It also results in highly optimized evaporator with no parasitic losses.

Table 1 shows the design parameters and physical characteristics of a case study. Also shown is an evaporator with same size tubes but no enhancement. Due to length restriction the maximum allowable tube length was limited to 16 ft (4877 mm). This restriction and the pressure drop limitation of less than 15 psi (1 bar) would have forced the design to 54” (1372 mm) shell diameter with 2600 plain surface tubes. As indicated in Table 1, this option would have had substantial effect on the cost and the ammonia charge. Maintenance and other running costs would have been additional expenses during the life of this evaporator.

During the design phase, several enhanced surface techniques were evaluated and it was finally decided to adopt a various-type enhanced tubes bundle. The depth of the bundle, two-pass arrangement (pressure drop limitation), LMTD, and the approach temperature directed the design towards a bundle with three different types of tubes. Figure 2 shows the tube hole layout with three designated sections I, II, and III each with similar type of tubes. The modeling process is cumbersome, proprietary, and cannot be disclosed. There is no model available in the open literature. Fourteen rows of the lower section had highly structured outside surface tubes with strong nucleate boiling characteristics and internal grooves (quantity 393) as shown in Fig. 3a. The middle section had the bulk of the tubes (quantity 649; 17 rows) with slightly wider gap structure on the outside in order to over come strong convective effects and similar internal enhancement as in the lower section tubes. This type of tube has shown good nucleate boiling behavior in the presence of strong convective forces (Fig. 3b). The decision to select appropriate tube in the upper section took the bulk of modeling and design time. Various calculations showed that the use of tubes similar to lower two sections would have aggravated the heat transfer due to vapor blanket phenomenon. Plain surface tubes could have been used, but tube side had glycol, which was in the final pass and would have cooled down, therefore, resulting in higher viscosity and further aggravating the problem. The calculations indicated that plain tubes would have not achieved the desired goal. Tubes with only internal enhancement in carbon steel or stainless steel are not readily available. Hence, it was decided to use twisted tape inserts in plain tubes as shown in Fig. 3c. After careful evaluation and modeling, a stainless steel tape with H/D = 5.5 (180° turn), width 0.5” (12.7 mm) and thickness .02” (0.51 mm) was selected. The clearance between the tape and the tube inside diameter was 0.06” (1.524 mm). It is worth noting that majority research work in the open literature insists on snug fit, however, this approach is only possible on the laboratory level and is not practical from the manufacturing and maintenance point of view. In fact a snug fit may have a negative effect on the thermal hydraulics as shown by Ayub and Al-Fahed (1993).

 

Click on Thumbnails to see Original Size

 

Table 1 Existing enhanced tube evaporator vs. plain tube flooded evaporator

Design capacity: 11,400,000 Btu/hr (3340 kW)

Refrigerant:       Ammonia @ +14°F (-10°C) saturated suction temperature

Process fluid:     25% wt/wt ethylene glycol brine

Process flow:      3500 gpm (221 l/s)

Process inlet:     +27°F (-2.78°C)

Process outlet:   +20°F (-6.67°C)

Developed By: