Thermosiphon Reboiler Shell/Tube Temps during Startup

[cwc1]

I am interested in how others are determining the shell and tube temperatures for the expansion joint analysis for vertical fixed-tubesheet thermosiphons, especially for the cold startup case in which the heating medium (steam) is introduced rapidly into the shell. In this case, the column sump and the reboiler tubes are full of subcooled (ambient) liquid. We have thermal analysis software that can produce a steady-state analysis including shell and tube Mean Metal Temps (MMTs), but the program expects that the fluid will be saturated at the free surface in the sump, and so it reassigns the temperature to equate to saturated conditions. If one simply assigns the tube temperature to equal the cold process temperature inside the tubes, it seems that the result could be a large degree of conservativism, since the tube temperature will actually be somewhere in between the stream temperatures. Then there is the time-dependent nature of the problem, with no circulation in the tubes initially, but then increasing circulation along with changing heat transfer conditions that would tend to shift the tube temperature more in favor of the tubeside fluid, so the worst case may not be at the instant the steam is applied. What approaches have been used or would be recommended? Has anyone done a full-blown CFD/FEA, and can any generalizations be made as to how the quick, cold startup case tends to compare to the operating cases in its severity? Thanks.

 

[chicopee]

I don't know anything about vertical fixed-tubesheet thermosiphons, however, on thing that I know is that it will take several hours for utility power boiler to go full steam; I think that the number is 100dF per hour during the warm up period of utility power boilers, however, I would have to double check that rate of temperature increase.
With vertical fixed-tubesheet thermosiphons, are these boilers anything like the old style vertical firetube boilers?

 

[cwc1]

The various styles of reboilers may be seen at

http://www.thermopedia.com/content/1078/

. Normally when a reboiler is started up the plant steam header is already charged, and the steam is introduced into the reboiler shell in a controlled manner through a control valve. There is potential, however, for rapid introduction of steam because of equipment or control system failure, or operator error. The need to analyze upset cases is pointed out by various industry references (e.g. API 660 - Shell and Tube Exchangers, Process Industry Practices, etc.), but how to convert the stream temperatures into the Mean Metal Temperature of the tubes is unclear to me for this transient case. Intuitively it may seem that the shell would heat much more rapidly than the tubes, because the tubes contain cold liquid. However, the liquid contained in the tubes is a fraction of the weight of the tube itself (for example, 12 gage, 1" diameter tubes containing organics), and with the shell being 3 times or more the thickness of the tubes, the thermal masses seem roughly equivalent in the static case. And even when circulation starts, the heat transfer coefficients for slow-moving subcooled organics will be very low, on the order of a few percent of the condensing coefficient on the outside of the tubes. So I am not so sure that the tube temperature really lags the shell temperature significantly. It would be possible to use the heat exchanger modelling software in normal (i.e. non-thermosiphon mode) and step through the transient, manually controlling the tubeside flow and balancing the pressure drop against the driving head from density difference, similar to how the program solves the thermosiphon case at saturated/boiling conditions. Before doing this I thought I would see how anyone else is handling it...thanks.

 

[chicopee]

These systems are outside my scope, however, may I suggest that you research the history of failures with these reboilers and you may discover clues to your problem. I don't know if the National Board of fired and unfired PV's have compiled accidents with these units but I know that this organization will describe accidental events on boilers and pressure vessels.