Externally flushed pumps

No," Zip answered. "Those were external flush pumps. Wouldn t want to get tar between them seal faces. Those pumps were flushed with diesel oil or light gas oil. On that kind of pump, the operator needs to adjust the seal flush pressure a few psi above the pump suction pressure. Best to have a little needle valve and a rotometer. Also, you gotta have a pressure gauge after the needle valve on the seal flush line."... 

That surely is a problem. That s why on external flush pumps, you gotta be real careful not to use too much seal flush pressure. Also, the seal flush liquid gotta be a kind that won t ruin the product being pumped, if it contaminates a little."... 

For externally flushed pumps, this is not a problem as long as the seal flush is commissioned before the pump is switched on. Which is a problem. 

When a pump is not in operation, it is not necessary to have the seal flush flowing. Obviously for self-flushed pumps, this is not a problem, as the seal flush flow stops with the pump. But for externally flushed pumps, there is no sense leaving the seal flush running. For example, let s say I m operating a Delayed Coker or Visbreaker charge pump. The seal flush flow for such a pump, all of which flows back into the pump suction when the pump is idle, would be one or two... 

As the inner seal for an externally flushed pump, such as a black oil (i.e., asphalt, resid), pump will continue to have the same loss of seal flush material to the process stream through the throat bushing with the conventional double mechanical seal as when the external seal is an seal, there is no real process incentive for use of the more costly seal. 

Referring to Fig. 37.1, in addition to valve 1, the valves shown as valves 2 and 3 must also be added. Without valves 2 and 3, the complete seal flush line cannot be filled with liquid. For externally flushed pumps, valves 2 and 3 are not required. 

This environmental control may also dilute the pumped product. If the external flush is water, it may be necessary to evaporate it later in a costly posterior process. 

To fix this problem in Port Arthur, I connected an external source of seal flush water to the pump, from a nearby washwater station. Pumps which have subatmospheric suction pressures, and which are not in continuous service, should not be self-flushed pumps. They should have an external source of seal-flush material connected to the mechanical seal. 

Horizontal pumps Single with external flush Extended seal life Low seal cost Pump can be run dry if flow is lost Seal water piping required Caustic and water will intermix Freeze protection of seal water piping may be required... 

For an externally seal-flushed pump, open the seal flush flow and adjust the seal flush pressure to the manufacturer s recommended pressure above the suction pressure. If this data is not available, use the pump suction pressure plus 10 percent of the pump s normal differential pressure. The seal flush pressure should not be close to the discharge pressure. This will damage the seal. 

When a mechanical seal is specified, it shall be located at the cover plate to seal the vapor in the supply tank or vessel. Mechanical seals normally seal vapor however, they shall be designed to operate in liquid in the event of tank or vessel overfilling. With purchaser approval, the vendor may supply a seal pressurized by flush from the pump discharge or from an external source. External seal flush must be compatible with the liquid being pumped. The seal chamber shall have provisions for a high point vent. 

The actuation chamber is made from stainless steel. It houses the canister within its adapter sleeve during dosing. A smooth surface provides a face seal between canister valve stem tip and sample injection passage. The close tolerance between the external wall of the adapter and the internal wall of the actuation chamber induces a partial vacuum which creates a controlled pumping action causing sample methanol to flush the injection passage after each dose. A replaceable 16 gauge needle fits onto a standard luer tapered tip. 

The checkers isolated the metal chlorides in the following manner. After the bomb had been opened and the top half of the reaction tube cut off, the bottom half of the tube was transferred immediately to a container similar to that shown in Fig. 20 but without a stopcock. This assembly was quickly transferred (2 to 3 minutes) to a large dry-box and opened, and the contents in the tube were washed with carbon tetrachloride onto a sintered-glass funnel and suction flask assembly connected to a suction system external to the dry-box. In this operation and subsequent washing with carbon tetrachloride, the chlorides were not drawn completely dry. After washing, the carbon tetrachloride-wet crystals were transferred to a smaller version of the container shown in Fig. 20 and pumped dry on a conventional vacuum line. The dried product and container (under vacuum) were then transferred to a smaller dry-box, where samples were taken for analysis. Both dry-boxes were kept flushed with PiOio-dried, purified nitrogen. 

Weter Leekage When water from external sources enters the process side of the evaporator, thermal efficiency decreases since such leaks dilute the product and increase the amount of water to be evaporated. Possible sources of water leakage into the process are corroded heat-exchange surfaces, pump seal flushes, barometric condenser water flooding, condenser flushes, and leaking of the water valves used to clean out the evaporator for startup and shutdown. 

The dynamic seal operates only when the pump is rotating at a sufficient speed. When the pump is stationary, the dynamic seal ceases to perform and liquid may leak through the stuffing box, unless an additional stationary seal is provided or external water at sufficient pressure is flushing the gland. 

With the external source of seal flush, the pump no longer lost its prime when shut down. However, 1 then noticed an interesting phenomenon. The pump would continue to run for about an hour after it stopped raining. It produced a discharge pressure of about 100 psig. Then, quite suddenly, the discharge pressure would drop to 70 psig and the amp load on the motor driver would slip from 20 to 14 amps for its last few minutes of operation. 

In black oil service. I ve seen asphalt plant operators wait a few minutes after a pump is shut down before they block-in the external seal flush flow. Of course, for autostart pumps, leaving the external seal flush closed is not an acceptable option. But, in the above example, maintain the autostart option is costing 2 million a year in degraded product yields. I suppose that s the justification for Nitrogen Barrier Seals, as described below. 

Referring to Fig. 37.1, note that the seal flush liquid enters the pump case between the rotating face (for internal assembly seals) and the back of the pump s impeller. The idea is for the relatively small flow of the seal flush liquid to flow back into the suction of the pump, and thus be returned to the process steam. (Incidentally, for external assembly seals, the fixed face is closer to the impeller. See reference 2 for details on internal vs. external assembly seals.)... 

---