Subject: Energy reduction
Date: Mon, 18 Dec 2000 16:54:12
Just a quick question, we have a naphtha stabilizer being revamped to produce lower RVP gasoline. One change is the addition of a second reboiler to get extra heat into the tower. The new exchanger uses steam to heat the bottoms. The existing exchanger is heat-integrated with a FCC main fractionator pumparound. When low RVP is not required (winter season) will the stabilizer be able to operate with only the pumparound reboiler in service? Can we shut the steam reboiler down when the extra duty is not required?
The reason I raised this question is that during winter, refinery steam consumption is higher and condensate return to the boilers can be a restriction due to the pumps being under-rated (we have improved our condensate recovery too much!). Since we will not require the low RVP during winter (at least not to start with) then we should try to minimize the steam usage during this time.
L., Pacific-Asia Refinery
Subject: Energy reduction
Date: Mon, 18 Dec 2000
The summary briefly covers the results, so you can get the results fast, then a detailed background will follow that you can look at in any level of detail you want.
To get to the point, the real question is if the piping and hydraulics would permit operation without heat input into the new reboiler. In summary, for your case, the reboiler hydraulic loop is tight if the reboiler has no heat input. The reboiler may require a small amount of heat input to vaporize the return stream to increase the heat available for flow. As an alternative, the line to the new reboiler could be increased one size along with the tower draw nozzle and the new reboiler inlet nozzle.
Modification to allow the steam side of the reboiler to run in either steam-chest pressure control or in flooded condenser control modes is recommended. This would allow operation at very low steam rates if necessary.
If the existing heat input can achieve a sufficient removal of light material to meet the future winter RVP requirements, obviously, the tower can meet the service requirements with the current duty. The real question is the lies in understanding the details of the reboiler hydraulic loops.
The question of taking the new reboiler out of service really brings up a discussion of how circulation goes through a reboiler system. Both of the reboilers on in your future stabilizer configuration are once-through reboilers. Figure 1 shows the configuration, approximately to scale, of the stabilizer and the new reboiler. The driving force for the flow on the process side of the new (first) reboiler is:
where the forces are shown on Figure 2. Critical in this is that the height difference between H1 and H2 is very small (around 18 inches, 45 cm).
With approximately 50% of the duty in each reboiler, the heat duty in the first reboiler vaporizes a considerable amount of the liquid. This reduces the average density2 term and provides DP for the pipe inlet losses and the reboiler losses.
With no duty in the reboiler, the DP losses in the reboiler and the exit nozzle to the tower are very low. The major pressure loss is in the inlet pipe to the exchanger, DPpipein. The 18 inches (45 cm) of head available are a close fit with the pressure losses through the system with no vaporization. If you want to run the system with no vaporization, moving up one line size in pipe to the exchanger, and exchanger inlet nozzle is recommended. Off course, to make this effective, the exit nozzle from the shell needs to increase one size as well.
For operating flexibility and to assure operation in all conditions, having the capability of running the exchanger with a low-heat-input is a good idea. In some circumstances, heat losses in the external loop through the out-of-service exchanger can change the liquid density enough to stop the liquid from flowing. This is one of the two reasons why having low duty capability on the exchanger should be included in the design.
The existing reboiler uses a recirculating design. The revamp switches the existing reboiler to a once-through design. One major reason to use a recirculating design is keep the vaporization percentage in the reboiler low. At high percent vaporization, vapor formed on the tube surface pushes liquid away from the tubes. This reduces the heat transfer effectiveness of the exchanger. By recycling liquid, recirculating thermosyphon exchangers can operate at lower percent vaporization.
In the no-steam to first reboiler operation, the vaporization percent in the second reboiler will be very high. This may induce the problem of vapor blanketing inside the tubes. Prediction of this behavior is very difficult. One mitigating factor is your case is that vapor blanketing inside tubes is less of a problem than vapor blanketing on the shell-side of the exchanger.
This is the second reason having the capability to operate the first exchanger at low heat loads is important.
This discussion covers the very specific case of having two reboilers on a column. The requirements for low duty control with one steam reboiler (that must meet process requirements) are different, so be cautious in applying the recommendations made here to other systems.
Steam control with no duty required is simple. Close the valves and let the exchanger sit. The control problem with a very low duty is more complex. For proper operation of the unit, facilities to handle low duties should be included.
Two main methods for utility control are available for controlling the heat input on the steam side. Exchanger performance is summarized by:
Duty in the reboiler must be controlled by manipulating either U, A or LMTDc. Once built, A (area) is set. Either U (overall heat transfer coefficient) or LMTDc (corrected log mean temperature difference) is manipulated.
Manipulating U involves changing a steam versus condensate level in the reboiler. The amount of surface area exposed to a high U action (condensing) area versus a low U action (subcooling condensate) sets exchanger performance. Manipulating T involves changing the steam pressure. A variable pressure drop upstream of the exchanger changes the condensation pressure on the steam side. Figure 1 shows the tower with the variable steam-side pressure control scheme. In comparison, Figure 3 shows an alternate with level control of condensate in the condenser.
While the steam-chest pressure control method is preferred because of its near-symmetrical control response to heat input demand changes and its speed of response, it does have some disadvantages. The disadvantages are:
The first two are inherent in the process. Economic tradeoffs for the superior control versus the extra equipment cost can be difficult to quantify, however, the variable steam pressure control is well worth the extra cost.
The turndown control issue is more complicated: but it does have several solutions. First, the problem, then the solutions: the operating pressure for the steam side of the exchanger lies between the supply pressure of the upstream steam system and the pressure required to pressure the condensate out of the system (of course, a condensate pump can be used, but this is rare). If the pressure is reduced to the minimum required to force the condensate out, and duty is still too high, then the exchanger must be partially flooded.
The same problem occurs if the exchanger is highly over-surfaced due to using large fouling factors during the design. This occurs at start-up conditions for services that are designed to highly fouled end-of-run conditions. The same solutions are possible.
Another problems for the scheme shown are the need to have an alternate control method for the tower heat input. With modern DCS systems, reconfiguring the control scheme is straightforward.
The solution here is to run the first exchanger in a fixed duty mode. The basic idea is to break open the control loop for steam flow to the exchanger. Open the steam-supply valve wide open, put the outlet condensate valve on manual flow control and let the exchanger flood to whatever level it equalizes on (Figure 4).
Such low levels of condensate may be desired that using the regular control valve may not be effective. A hand-operated control valve placed in parallel with the existing control valve may give the best control (Figure 5). The ability to add this as required without shutting down is recommended (include tie-ins in base revamp).