Oct-2014

Troubleshooting steam ejectors

Some causes and practical solutions concerning problems with refinery steam vacuum ejectors

NORMAN LIEBERMAN
Process Improvement Engineering

Article Summary

The majority of vacuum ejectors in refinery services do not perform as per their design. Steam vacuum ejectors, or jets, are widely used in both condensing steam turbine surface condensers used to generate power and drive large compressors. Also, steam jets are employed in vacuum towers which are used to produce paving asphalt and lubrication oils.

Poor performance of a steam turbine surface condenser will typically increase the energy requirements to generate electricity by 10%. Poor performance of a steam jet in a refinery may cost the refiner $20 000 (US) per day in reduced gas oil recovery from visbreaker bottoms or from delayed coker feed. Degraded vacuum may preclude meeting the viscosity and flash point specification for the production of paving grade road asphalt or industrial fuel oil.

Many refiners are able to sustain a reasonable vacuum during the winter, but find that their vacuum breaks and their steam ejectors start to make a surging or hunting sound when summer arrives.

The problems that give rise to these malfunctions number in the hundreds. A book on this subject1 has been rightly criticised for omitting half of the common causes for vacuum system malfunctions. The following tabulation is far from complete, and only represents the most common malfunctions seen most recently.

Erosion of steam nozzle threads
The steam nozzle (see Figure 1) is typically machined from 316 (L) stainless steel. It is not generally subject to damage to its threads. The major problem is that the stainless nozzle screws into the carbon steel ejector backing plate. The female threads of the backing plate corrode-erode, apparently due to galvanic corrosion and/or erosion due to wet motive steam. The motive steam then partly by-passes the steam nozzle and blows into the mixing chamber. The moisture in the steam acts like an electrolyte. Superheated steam would reduce this galvanic corrosion problem.

Another common cause of steam nozzles becoming loose because of erosion of their threads is because the nozzles were never installed properly. Either the threads were dirty or the nozzle was not screwed in tightly. This reportedly happens when the steam nozzles are replaced in the field. However, we have noted this problem on original ejectors too, with no history of nozzle removal. Whatever the cause, a small amount of leakage around the threads will escalate due to steam erosion. So, if steam nozzles are removed for inspection, clean off all the threads and reinstall the nozzles tightly using Teflon tape.

If you notice an improvement in vacuum by reducing the motive steam pressure by 20% below the design motive steam pressure, eroded threads may be the problem. We have temporarily fixed this failure with lots of Teflon tape. Why do the ejector manufacturers not use a stainless steel backing plate, and weld the steam nozzle to the ejector and avoid this problem? Actually, several refiners do follow this useful practice.

Plugging of steam nozzles
We have written about this subject and published photos of partly plugged steam nozzles extracted from a steam turbine surface condenser.2 The problem here is silicates and other hardness deposits in the motive steam supply. This is a consequence of poor level control in upstream waste heat steam boilers. This is such a common problem that the manufacturers typically provide a small, screwed, clean-out plug in the backing plate of the ejector (see Figure 1). As most vacuum systems have two or three parallel sets of jets, one could isolate one jet and easily clean out the silicates while still on-stream.

Incidentally, if one steam ejector with a relatively clean nozzle is operating in parallel with a jet with a fouled nozzle, the good jet will extract motive steam out of the fouled ejector and greatly degrade the efficiency and capacity of the clean ejector.

Distance between steam nozzle and diffuser throat set wrong
The distance between the steam nozzle and the inlet to the diffuser (see Figure 1) is adjustable. There are spacers that are installed between the ejector backing plate and the steam nozzle that are used to make this adjustment which is supposed to be done in the factory. This adjustment is, at least partly, a function of the motive steam pressure. On several troubleshooting projects, the author, having failed to correct the ejector malfunction, was told by the vendor that the steam nozzle position had been incorrectly set in the factory for the current motive steam pressure and temperature.

Moisture content of steam
The variable that provides the energy to compress the gas to the ejector is the kinetic energy of the motive steam. This kinetic energy is not derived from the motive steam pressure but from the motive steam’s enthalpy or heat content. If the steam is wet, the moisture flashes to steam in the steam nozzle. The resulting conversion of the sensible heat of the steam to latent heat of evaporation of the entrained water reduces the speed or velocity of the steam entering the diffuser. In one recent case, wrapping 25 ft of a two-inch steam supply line with loose insulation improved the vacuum from 54 to 49 mm Hg.

On the other hand, highly superheated steam is also marginally bad for the ejector’s performance. Hence, steam condensate is injected through a ‘desuperheating station’. However, it is not uncommon for the temperature controller to be set a few degrees below the saturation temperature of the motive steam. The author has seen this done at two refineries recently. The effect, of course, is to inject uncontrolled water into the motive steam. Auto-refrigeration of the exterior of the converging section of the ejector will be observed. Also, there will be erosion of the carbon steel threads where the stainless steel steam nozzle screws into the ejector body as a consequence of the water in the steam. In one refinery, correcting this problem improved vacuum from 28 to 15 mm Hg.

Air leaks that cause nitrogen in the off-gas to exceed 10-20%
The easy way to find significant air leaks is to look for moisture condensing on the ejector’s body, barometric drain legs, or flanges. Air, as it is drawn into the diffuser or flange, expands and auto-refrigerates. In areas of high humidity, atmospheric moisture will condense out on the exterior of chilled metal surfaces. The quite common solution for on-stream repair is duct tape. This is a permanent and inexpensive repair technique. Air leaks in a seal leg will cause poor drainage through the affected leg and condensate back-up.

High concentrations of CO2, as opposed to oxygen, in the vacuum tower off-gas is an indication of an air leak in the line connecting the fired heater to the vacuum tower –obviously a potentially dangerous situation which requires careful inspection of the heater’s ‘transfer line’. Imagine what may (and has) happened if the vacuum is suddenly lost and the pressure in the transfer line goes positive. Several catastrophic fires have resulted.