[pullquote]
Most compressors have suction drums or knockout pots upstream for protection against liquid slugs entering the compressor feed. The droplet size being removed rarely is explicitly stated. The equivalent size normally is buried in a sizing coefficient relating drum diameter to the vapor rate for the drum. Depending upon the sizing criteria used, most drums are designed to remove liquid droplets whose diameter exceeds some size between 100µ and 150µ (0.0039–0.0059 in.).
A client had hydrogenation reactor effluent going to a separator drum. The overhead gas from the separator was hydrogen; it went directly to the suction line of a dedicated reciprocating compressor.
Inspection after a shutdown showed significant solid deposits in the compressor suction line. Laboratory work identified the deposits as solidified reactor effluent. The drum feed was a liquid/vapor mixture at normal operating conditions of 260°F but the liquid solidified around 150°F.
The concern was that liquid was entraining — such entrainment could damage the compressor. Unfortunately, simply monitoring compressor vibration may not suffice when evaluating entrainment. In multiple incidents, liquid entrainment has resulted in compressor operation with very low vibration levels; the compressor operated successfully for an extended period but then catastrophically failed.
You should monitor compressor vibration and take appropriate corrective action when vibration levels are high. However, if you suspect entrainment, remember that low vibration levels are a poor indicator that entrainment isn’t causing damage.
The concern about entrainment prompted the plant to consider adding a knockout drum downstream of the separator. This is common practice in hydrogenation reactor systems. Indeed, roughly 80% of hydrogenation reactors with operating conditions similar to those at this plant have an effluent separator for bulk vapor/liquid separation followed by an additional knockout pot to protect the compressor suction. The knockout pot normally has a trip that shuts down the compressor on high liquid level.
The knockout pot only gets a small quantity of liquid in its feed. The pot holds a liquid level but removal of liquid only happens sporadically, i.e., the liquid out of the pot is a NNF (normally no flow) stream. If the pot requires frequent draining, this indicates the upstream separator is undersized or having problems.
Of the about 20% of hydrogenation reactors that don’t use such a setup, about half directly feed the recycle compressor from the separator drum without using a knockout pot. The other half have two knockout pots in series downstream of the separator — despite the fact that, generally, if a single knockout pot doesn’t remove enough small liquid droplets, a second knockout pot of the same size won’t do any better. Other combinations, such as a knockout pot and coalescer or filter, are possible but rare.
The plant had purchased a used vessel to separate the reactor effluent; it was supposed to be oversized for this service. In fact, the standard GPSA method showed that liquid droplets smaller than 51µ would settle in this drum. The drum had a vapor velocity far lower than a typical good-practice drum needs in this service. The vessel amply should have done the job of liquid removal.
Heat loss through the pipe was the major reason for the deposit buildup. A brief inspection showed pipe temperature was low enough to cause any liquid that wetted the pipe surface to solidify. (By the way, we identified two possible sources of downstream liquid.) The fix to prevent solidification is simple, insulate the lines better.
Even if the drum works perfectly, some liquid can go downstream. Over time, enough small droplets, <51µ in this case, can coat the line and cause significant deposits to accumulate gradually. However, because droplet sizes and total quantities reaching the compressor are small, this liquid shouldn’t result in compressor damage.
The other possible culprit was the inlet line to the separator because it can affect separator performance. Expectations for what a good inlet line looks like have changed over the years. Separator guidelines in the 4th edition of the GPSA Engineering Data Book don’t even discuss the inlet line but the current, 14th edition includes specific advice for the inlet piping to the vessel.
The plant’s repurposed vessel had a particularly bad inlet piping configuration. The inlet pipe enters the vessel vertically and points toward the vapor outlet. Even though the distance between the two is large and inlet velocities are low, this isn’t a good practice and one always to avoid. While insulating the line to the compressor will prevent solid deposits forming, liquid entrainment may be occurring. The data on this aren’t clear.
The ideal solution requires both insulation and modifications to the separator drum inlet piping.
