Apr-2015

Reducing vacuum tower pressure

To create a better vacuum the vapour load to the ejector has to be reduced

NORMAN LIEBERMAN
Process Improvement Engineering

Article Summary

In a previous issue of PTQ,¹ we discussed several troubleshooting opportunities in correcting vacuum system ejector and condenser malfunctions. In this article, several field-proven techniques to improve vacuum by process changes are detailed. While none of these techniques are common, they do work and are not proprietary.

Slipping HVGO into LVGO pumparound
The purpose of the LVGO pumparound is to minimise the vacuum tower overhead vapour flow to the vacuum system. The vapour flow from the top of the vacuum tower is proportional to:

Absorption factor = L/VK = L/V • PT/VPi                            (1)

where
V = Moles of vapour
L = Moles of liquid
PT = Pressure
VPi = Vapour pressure of the ith component

The bigger the absorption factor, the lower the amount of vapour to the ejectors. One way to increase the absorption factor is a colder tower top temperature. The other way is to increase L in the above equation. This can be done in two ways (see Figure 1):

1.    Reduce the HVGO pumparound heat duty and increase the LVGO pumparound heat duty. L in Equation 1 is not the LVGO pumparound rate, but the net LVGO product rate. However, reducing the HVGO pumparound heat duty will also, most likely, diminish the crude preheat. Also, the increase in the LVGO pumparound duty may raise the vacuum tower top temperature if the LVGO pumparound rate is limited. The resulting increase in the top temperature can offset the beneficial effect of the increased LVGO circulation rate. That is, the vapour pressure of the components (see Equation 1) may increase faster than the moles of liquid. Our field experiments have indeed shown that sometimes raising the tower top temperature, by reducing the HVGO circulation rate, improves vacuum, and sometimes it degrades vacuum.

2.    Our second method of increasing L (the LVGO net product rate) in Equation 1 allows for an increase in vacuum, but without loss of HVGO pumparound heat duty and crude preheat. However, this method will not be possible if the LVGO product is flowing to a hydrocracker and the HVGO product is being used as FCC unit feed. However, if this is not the case, and both LVGO and HVGO products are combined, then we have directed a small portion of the HVGO product to the LVGO pumparound return. We have observed a few mm Hg reduction in the vacuum tower top pressure as a result. The amount of HVGO diverted in this manner was not metered.

Increasing tower top temperature
We have observed that on most vacuum towers with a pre- condenser (see Figure 2), raising the tower top temperature up to a point improves vacuum.² The reason for this again has to do with Equation 1. Raising the top temperature increases the moles of heavy naphtha and kerosene (180-240°C boiling range hydrocarbon components) distilled overhead. This increases L, the moles of absorber oil, in the pre-condenser, and thus reduces the moles of vapour (V) flowing to the first stage ejector.

However, we have also observed that if the vacuum tower top temperature is raised too far then the pre-condenser outlet temperature will also begin to increase excessively. The net effect is similar to reducing the HVGO pumparound duty too much. That is, the moles of vapour escaping from the pre-condenser will increase as the vacuum tower top temperature is increased too much.

As Figure 2 shows, we have injected a naphtha stream as an absorption oil (akin to the sponge or lean oil in an absorber) to the inlet of the pre-condenser, to increase L in Equation 1. A word of caution: the first time we did this, we unfortunately used a light naphtha product, rather than heavy naphtha or kerosene. The light naphtha injection led to a loss in vacuum, as it largely vaporised in the pre-condenser. To avoid the ire of our clients, we now calculate the vapour-liquid equilibrium flash in the pre-condenser before, rather than after, implementing process changes in the field.

The heavy naphtha injected will typically be recycled through the condensate seal drum along with the steam condensate through the desalter (with the wash water) and then through the crude unit and the naphtha stabiliser and splitter. Thus, there is a price to be paid for improving the vacuum with this method, unless the seal drum condensate can be separated from the steam condensate and directed to a downstream hydrotreater without prior prefractionation.

H2S extraction from vacuum tower off-gas
We have often sampled vacuum tower off-gas from the seal drum. Assuming that:
• Air leaks are small (less than 5% nitrogen in the off-gas sample)
• Vacuum tower feed is well stripped for light ends removal (less than 6% propane in the off-gas sample)
• High sulphur crude is being run (1.5-2% sulphur).

Then, the amount of H2S in the off-gas (on a dry basis) will typically be 30-40 mole%. (Caution: H2S at a concentration of 0.1 mole% is quite fatal to breathe, so fresh air equipment is advisable when obtaining this sample.)
Extraction of this H2S with an amine (MDEA) would then reduce the vapour load to the downstream jet by a very large amount, somewhat greater than 30-40%, as the H2S has a greater molecular weight than steam or air. This can be done as shown in Figure 3. The H2S scrubber depicted would be identical to the now obsolete barometric condenser design.

The author has never actually used amine as an H2S absorption agent, as suggested here. However, he has used NH3 in the same manner as the proposed use of the amine. The NH3 connection was intended as an HCl neutraliser. By using the NH3 at a rate hundreds of times above that which was intended for HCl neutralisation, the vacuum was vastly improved.

Unfortunately, this benefit only lasted for half a day. That is, until the supply of neutralising NH3 was exhausted. As the client summarised, “this was a technical success, but an economic fiasco”.

Of course, the off-gas has to be scrubbed with amine regardless, to extract the H2S, so the use of amine, as opposed to NH3, ought to be an economically viable project.

A safety note: CO2 will also be extracted by the amine. But CO2 in the vacuum tower off-gas is an indication of an air leak in the vacuum heater transfer line. This is a serious safety issue because, when vacuum is lost, 400°C resid can, and has, blow out of this leak, auto-ignite, and burn down the vacuum tower. This happened to a vacuum tower that the author revamped in a plant in the Baltics several years ago. CO, on the other hand, is just an indication of the thermal degradation of naphthenic acids, which is unavoidable and is not a sign of any avoidable problem.

Cracked gas evolution in boot
On some refinery vacuum towers, especially those that lack a boot quench, we have measured by varying the boot level that the non-condensable vapour load due to thermal cracking in the boot may be 30-40% of the total non-condensables.