1 reply to this topic

[unvecchietto]

Hi everyone,

I am designing a total condenser operating under full vacuum conditions. The process involves steam entering the shell side at 165°C and leaving as condensate. The tube side uses a process gas as a coolant (80,000 kg/h) in a U-tube bundle (2 passes). The exchanger is relatively "short and fat" (approx. 1000 mm length x 740 mm diameter), in BXU configuration.
The main challenge is the very low steam flow (5,000 kg/h) relative to the s&t size, which creates significant hydraulic distribution and duty control issues.
I have tested these configurations
1. BXU: Pressure drop is concentrated 100% in the inlet nozzle. The bundle is hydraulically "transparent" (approx 0.3 kPa in the cross). This clearly violates the rule of thumb of bundle pressure drop > 66.7%of the total to ensure flow uniformity. Without this resistance, I am concerned about a localized "rain effect" directly under the nozzle, leaving the rest of the 1-meter bundle ineffective
2. BJ21U: Similar results. Despite the divided flow, the longitudinal velocity is too low to "activate" the bundle hydraulically.
3. TEMA BGU :This configuration performs better regarding distribution because the longitudinal baffle and the required path force a higher internal resistance.

The system uses Flooded Surface Control (varying the condensate level to adjust active surface area).
For a BGU design, the longitudinal baffle must be continuously welded to the shell. Does this create excessive thermal stress risks in vacuum service, or is it a non-issue for a 1-meter shell?
If I still use a Type X , is the combination of an Annular Distributor (Vapor Belt) and a perforated Distribution Plate (min. 20% open area) considered reliable enough to satisfy the 2/3 Rule and distribute vapor over such a short bundle?
Given that the BGU uses segmental baffles (even with 12 mm drainage notches), is the condensate level control as linear and stable as in a Type X, which provides a perfectly flat free surface?
In the worst-case clean turndown, what measures would you recommend to mitigate water hammer risks in the presence of low-pressure vapor?

I would appreciate insights from anyone who has dealt with this type.of vacuum condensers and the trade-offs between cross-flow and split-flow hydraulics. Is type G suitable for this case? I have to evaluate the submerged area during the turndown? Would you still make a Type X or Type J with a vapor belt and perforated plate instead of a Type G? Although with vapor belts the situation in terms of DP would not improve much?

Thanks a lot.

[Pilesar]

If the purpose of the exchanger is to condense steam on the shell side under vacuum conditions, then the exchanger should have extremely low pressure drop on the shell side. There are no hydraulic issues caused by "too low pressure drop." The 2/3 rule does not apply. Reduce your pressure drop wherever you can by as much as you can. Use large inlet nozzles and minimal tube baffles. The steam will distribute itself. The problems will be caused by inerts in the steam which will be difficult to remove from the exchanger. The steam will travel rapidly to the tubes as that is where the condensation removes vapor. Any inerts will be carried along with the steam to the tubes where the inerts will build up in the vicinity of the tubes which will reduce the partial pressure of the steam in the vapor and impede performance. Another major issue is how to get the condensate quickly away from the tube surface so that the steam can reach the heat transfer surface. Vacuum steam condensers should be specially designed for their specific trouble areas. When designed by experts, they look quite different from the "usual" shell and tube exchangers. Some features I put in my designs: leaving paths of "missing" tubes in the bundle to make it easier for steam to travel to the bundle interior, venting inerts near the level of the condensate since air is heavier molecular weight than water, forcing the vent path to travel across tubes by use of interior bundle baffling to reduce steam loss.