Tower pressure controls

I well remember one pentane-hexane splitter in Toronto. The tower simply could not make a decent split, regardless of the feed or reflux rate selected. The tower-top pressure was swinging between 12 and 20 psig. The flooded condenser pressure control valve, shown in Fig. 3.1, was operating between 5 and 15 percent open, and hence it was responding in a nonlinear fashion (most control valves work properly only at 20 to 75 percent open). The problem may be explained as follows. 

Another reason to raise tower pressure is to permit higher reflux rates. If the pressure controller in Fig. 3.1 is set too low, then during hot weather, when condenser capacity becomes marginal, the level in the reflux drum will be lost. If we then raise the pressure set point, the drum will refill—but why ... 

Regardless of the method elected, the principal concept of tower pressure control is the same. We control the pressure in the reflux drum by manipulating the temperature in the reflux drum. The tower pressure then floats on the reflux drum pressure. To lower the tower pressure, we must first cool the reflux drum. This cuts the vapor pressure of the liquid in the reflux drum. 

The oldest, most direct method of pressure control is throttling on the cooling-water supply. This scheme is shown in Fig. 13.5. Closing the water valve to the tube side of the condenser increases the condenser outlet temperature. This makes the reflux drum hotter. The hotter liquid in the reflux drum creates a higher vapor pressure. The higher pressure in the reflux drum increases the pressure in the tower. The tower pressure is the pressure in the reflux drum, plus the pressure drop through the condenser. 

Figure 13.5 Tower pressure control using cooling-water throttling.

Hot-vapor bypass pressure control. A more modern way of controlling a tower s pressure is shown in Fig. 13.6. This is the hot-vapor bypass method. When the control valve on the vapor bypass line opens, hot vapors flow directly into the reflux drum. These vapors are now bypassing the condenser. The hot vapors must condense in the reflux drum. This is because there are no vapors vented from the reflux drum. So, at equilibrium, the hot vapors must condense to a liquid on entering the reflux drum. They have no other place to go. 

As the hot-vapor bypass valve opens, the condensate level in the shell side of the condenser increases to produce cooler, subcooled liquid. This reduces the surface area of the condenser exposed to the saturated vapor. To condense this vapor, with a smaller heat-transfer area, the pressure of condensation must increase. This, in turn, raises the tower pressure. This then is how opening the hot bypass pressure-control valve increases the tower pressure. 

In general, flooded condenser pressure control is the preferred method to control a tower s pressure. This is so because it is simpler and cheaper than hot-vapor bypass pressure control. Also, the potential problem of a leaking hot-vapor bypass control valve cannot occur. Many thousands of hot-vapor bypass designs have eventually been converted—at no cost—to flooded condenser pressure control. 

If we normally have a situation in which noncondensable vapors appear in the reflux drum, then there is only one pressure-control option available. This is to place the tower pressure-control valve on the vapor off-gas as shown in Fig. 13.9. If we normally have noncondensable vapors in the condenser effluent, then the following problems we have been discussing do not exist ... 

Sometimes we see tower pressure control based on feeding a small amount of inert or natural gas into the reflux drum. This is bad. The natural gas dissolves in the overhead liquid product and typically flashes out of the product storage tanks. The correct way to control tower pressure in the absence of noncondensable vapors is to employ flooded condenser pressure control. If, for some external reason, a variable level in the reflux drum is required, then the correct design for tower pressure control is a hot-vapor bypass. 

Figure 3.14. The lower ends of fractionators, (a) Kettle reboiler. The heat source may be on TC of either of the two locations shown or on flow control, or on difference of pressure between key locations in the tower. Because of the built-in weir, no LC is needed. Less head room is needed than with the thermosiphon reboiler, (b) Thermosiphon reboiler. Compared with the kettle, the heat transfer coefficient is greater, the shorter residence time may prevent overheating of thermally sensitive materials, surface fouling will be less, and the smaller holdup of hot liquid is a safety precaution, (c) Forced circulation reboiler. High rate of heat transfer and a short residence time which is desirable with thermally sensitive materials are achieved, (d) Rate of supply of heat transfer medium is controlled by the difference in pressure between two key locations in the tower, (e) With the control valve in the condensate line, the rate of heat transfer is controlled by the amount of unflooded heat transfer surface present at any time, (f) Withdrawal on TC ensures that the product has the correct boiling point and presumably the correct composition. The LC on the steam supply ensures that the specified heat input is being maintained, (g) Cascade control The set point of the FC on the steam supply is adjusted by the TC to ensure constant temperature in the column, (h) Steam flow rate is controlled to ensure specified composition of the PF effluent. The composition may be measured directly or indirectly by measurement of some physical property such as vapor pressure, (i) The three-way valve in the hot oil heating supply prevents buildup of excessive pressure in case the flow to the reboiier is throttled substantially, (j) The three-way valve of case (i) is replaced by a two-way valve and a differential pressure controller. This method is more expensive but avoids use of the possibly troublesome three-way valve.

Packed towers may be operated above 90% of the flooding velocity when pressure controls are provided to maintain a fixed maximum pressure drop not to exceed the desired flooding velocity. Strigle suggests... 

---