Stahl Columns: An alternative to molecular sieves?
Case study and independent cost estimations lay out what may someday become the standard way to achieve deep dehydration of produced gas streams.
Jeffrey Weinfeld, Nathan Hatcher, Daryl Jensen, Ralph Weiland & Chris Villegas
Optimized Gas Treating, Inc.
Viewed : 2672
Gas dehydration is used to remove water vapor from gas streams for applications such as pipeline transportation and cryogenic processing. Absorption into glycols is the favored dehydration method for pipeline transportation, where typically a moisture content between 80 and 140 ppmv is required although 7 lb/MMSCF is the often-quoted gas-industry standard1.
But for deep dehydration applications such as LNG processing, where a moisture content of usually less than 0.1 ppmv is required, glycol dehydration for bulk water removal followed by adsorption onto molecular sieves is currently the most common method2. However, even though the molecular sieve process can achieve deep dehydration, it is economically expensive.
Disadvantages such as high pressure drop, maintenance due to bed changes and maintaining switching valves can amount to high costs. On the contrary, the glycol absorption process is economically favorable, but the inability to adequately strip the wet solvent with a reboiled column has heretofore prevented it from being used in really deep dehydration applications.
Deep water removal
In the glycol process, the achievable moisture content of a gas being dehydrated is almost entirely controlled by the dryness of the lean solvent (glycol) and the temperature of the wet inlet gas. On the solvent side, dried lean solvent is obtained by regeneration of the wet glycol. Concerning the controlling temperature, only a small solvent flow is needed to treat a large gas flow so the L/V ratio in a dehydration column is usually quite small.
The thermal mass of the solvent flow relative to the gas is therefore too small to greatly affect the gas temperature in most of the absorber. Thus, contrary to popular belief, most of the dehydration column is usually close to the temperature of the entering gas, not to the temperature of the lean solvent (except at the very top of the column where the gas rapidly cools or heats the glycol).3
From a process standpoint, the conventional reboiled regenerator has a cripplingly-serious, inherent weakness in the context of deep water removal. The dehydrating agent is saturated steam, but water is the very component that is desired to be removed from the solvent. There is no carrier or diluent for the removed water. In other words, there’s no place for the stripped water to go except into the already saturated steam.
The driving force for stripping out the water is the difference between the equilibrium and actual water content of the vapor. These quantities are very nearly equal throughout most of the column; thus, there is just little or no driving force for stripping water from the solvent when the vapor is already almost all water. This is the flaw in the solvent regeneration side of the process, but it can be overcome by providing a diluent gas.
To a very limited extent, this diluent is already provided by the gases that dissolved into the glycol in the dehydration column (usually at high operating pressure), but they are released near the top of the column where they are immediately swept out. They therefore do the least good because they have such a small volume of the column in which to operate. In any case, their concentrations are usually far too low to have a significant dilution effect.
They also may contain some of the very components having significant sales-gas value, so it is desirable to keep them with the sales gas. In addition, there may be components with serious environmental concerns if released into the atmosphere with the stripped water vapor. Incidentally, most glycol regenerators are refluxed. This provides no benefit to dehydration because putting some of the already stripped water back into the column is counterproductive.
Condensate is recycled to recapture glycol from the vapor via a water wash, not to enhance dehydration. The boiling of solvent in the reboiler is what is chiefly responsible for stripping water from the wet glycol. The column itself contributes very, very little. What small benefit it has is to a considerable extent, destroyed by returning reflux water to the top of the regenerator to recover glycol vapor before it escapes from the system.
At best, the reboiler is a single ideal stage of contact and the rest of the regeneration system, mostly the column, is functionally dormant as far as water removal is concerned. To activate the column itself requires use of a stripping gas to dilute the stripped water vapor and encourage further evaporation. This is the principle behind the Stahl column.
The focus in the following case study is on whether it is possible in principle to regenerate triethylene glycol (TEG) to a moisture level capable of drying methane to below 0.1 ppmv H2O, the generally accepted maximum moisture level recommended for gas entering the liquefaction section of an LNG train. It goes without saying that this is not possible using only a reboiled regenerator.
Our contention, substantiated via simulation, is that a Stahl column can enable reaching the same low water content while keeping temperatures below TEG’s decomposition limit. Recently, Carmody4 presented an interesting paper in which he suggested using the approach being described here. Without access to a mass transfer rate-based simulator, however, his analysis could not connect ideal stages to actual towers with real internals. Here we show that a TEG system alone can achieve dehydration very satisfactory for gas liquefaction in an LNG plant without using molecular sieves at all.
To test this idea, a plant was designed with specifications similar to existing TEG plants. A Stahl column was added to the regeneration section immediately below the reboiler of the conventional regenerator. A small slipstream, 0.08% of the fully dehydrated gas, was fed to the bottom of the Stahl column where it acts as very dry stripping gas. The contactor contains 10-m of MellapakPlus 452.Y, the stripper contained 3-m of 1-inch Metal Pall Rings, and the Stahl column contains 10-m of Mellapak 250X. None of these columns is particularly tall and all are well within reasonable flooding levels. The L/G ratios are typical for glycol dehydration units so an unusual hydraulic situation in any of the columns is not expected.
Add your rating:
Current Rating: 3