Introduction to LNG Processes
The refrigeration and liquefaction sections of any LNG project are very costly items of equipment. A number of cycles have been developed to achieve the low temperatures required. Brief descriptions of a small number of process routes will be give.
Some twelve countries were operating by 2003. The cycles employed are the classical cascade, the mixed refrigerant and the propane pre-cooled mixed refrigerant.
The process selected must ensure high onstream factors, reliability of equipment, flexibility and ease of operations and guaranteed capacities. Small differences in projected thermodynamic efficiencies are usually less than the uncertainties in equipment performance. A proper choice must be made between the use of well-proven technology and the acceptance of innovations designed to improve efficiencies and reliabilities.
In addition to the liquefaction equipment, a facility must have extensive gas purification equipment, designed to operate at high pressure, together with large compression machinery, special materials of construction, complex heat exchangers and large tankage.
see also our LNG and LPG Books section
Gas piped to a liquefaction plant from fields around 100 miles away is usually only given the minimum purification necessary at the well head. The feed gas may therefore contain water, carbon dioxide, hydrogen sulphide, higher hydrocarbons and other impurities. The line must be pigged regularly to prevent blockage and irregularities due to two-phase flow. The first-stage treatment will therefore comprises traps and facilities to collect liquids. Depending on ambient temperature, water content and pressure drop, glycol or methanol can be injected to prevent hydrate formation. In this case, glycol/methanol recovery involving fractionation will be required on the aqueous layer in the gas/liquid separator.
After reduction of liquid water, glycol and heavier hydrocarbons by simple gas/liquid separation, the gas is cooled by heat exchange to a few degrees below freezing.
At pipeline pressure, this results in further condensation, and more water and heavy hydrocarbons separate out in a knockout drum.
The next stage, gas sweetening, removes H2S and CO2. The reagents used can operate either by a reversible chemical reaction between the acid gas and the solvent or by the acid gas dissolving in the absorber liquid in preference to the other gas components. In the case of physical absorption, equilibrium concentrations of H2S and CO2 in the liquid are strictly proportional to the partial pressures of the gases. Reactive solvents, however, have absorption equilibria independent of the gas partial pressure.
Most solvents used for absorption of acid gases are non-or only partially selective. CO2, H2S and mercaptans are removed roughly in proportion to their original concentration in the gas. Typical chemically reactive solvents include aqueous solutions of most alkanolamines such as monoethanolamine (MEA), diethanolamine
(DEA), diglycolamine (DGA), di-isopropanolamine (Adip), triethanolamine (TEA) and anthraquinone disulphonic acid (Stretford solution). Apart from the latter, the acid gases are absorbed at near ambient temperature by the alkaline compound and are released by heating to near its boiling point. The Stretford solution, which also contains sodium vanadate, sodium carbonate and a trace of chelated iron, when blown with air, oxidises HS to elemental S which can be removed by filtration.
Another series of absorption solvents are based on potassium carbonate and act similarly to the alkanolamines. In the Benfield, Vetrocoke and Catacarb processes, the C02 reacts to form bicarbonate, which decomposes at elevated temperatures.
A similar reaction occurs with H2S, and here, various additives, often arsenates, assist H2S removal by forming thioarsenates which decompose into arsenates and elemental sulphur as in the Giammarco Vetrocoke process.
Physical absorbents for acid gases include anhydrous propylene carbonate (Fluor solvent), N-methyl-2-pyrrolidone (Purisol), the dimethyl ether of polyethylene glycol (Trigol). A disadvantage compared with chemical absorbents is a tendency to remove higher hydrocarbons from the gas. This can cause problems in Claus plant. The main disadvantage is the corrosive nature of absorbents and absorbent–acid compounds. Hybrid processes have been developed to alleviate the problem. An example is the sulfinol process which uses a mixture of the physical solvent sulpholane and chemical absorbents of the alkanolamine type.
Read Also What is LNG
The choice of LNG process for a particular application depends on:
- The original pressure of the gas—high pressure assists physical absorbtion
- The original concentration of the acid gases—in general, chemical absorbents have a higher absorption capacity
- The relative concentrations of the acid impurities—some solvents are more selective than others
The permissible residual concentration of the impurities
- The presence or absence of COS and CS2.
After sweetening, the gas is generally saturated with water which obviously has to be removed before liquefaction. Drying can be carried out by
1. Simple refrigeration: for instance in a turbo expander.
- Glycol dehydration: in a counter-current scrubber using di-, tri- or tetraethylene glycol at gas temperatures from 15 to 65 C. The saturated glycol is continuously regenerated by stripping at 200 °C. Gas dew points of −70 C can be achieved. Higher hydrocarbons are removed separately.
- Solid desiccant adsorption: using desiccants such as silica gel, alumina or
molecular sieves. The gas is passed at pressure through a packed adsorption vessel usually in a two-tower system where both hydrocarbons and water are removed. The towers are switched for regeneration by hot gas flow through the bed. The water and hydrocarbons can be recovered by cooling and condensation.
Hydrocarbon and water dew points of −50 to −70 °C can be achieved. The use of certain molecular sieves permits selective separation of water, hydrocarbons sulphur compounds and carbon dioxide. However, the adsorption capacities of the sieves limit these processes to final gas clean-up.
The level of impurity that can be tolerated in liquefaction plants LNG depends on the actual process chosen. Generally, the water content should be less than 1 ppm, carbon dioxide concentration 50–150 ppm, hydrogen sulphide could be as high as 30–50 ppm, but considerations of odour, corrosion and toxicity restrict it to a maximum of 3 ppm. Higher hydrocarbons could be removed by separation facilities in the liquefaction section of the plant, but removal prior to liquefaction is better.
LNG Liquefaction Processes
Refrigeration is based on the conversion of internal energy of a fluid into external work, and the second law of thermodynamics imposes a limit on the efficiency with which such a conversion can be carried out. To operate between two temperatures, e.g. the boiling point of a fluid at compressor exits pressure TB and ambient temperature TA. A fluid undergoing a closed cycle would absorb a minimum amount of mechanical energy as expressed as:
In order to transfer energy equivalent to Q from TB to TA, the efficiency term (TA − TB)/TB defines the ideal (reversible) energy required for refrigeration. In practice, the efficiency of actual processes falls seriously short of such values. To convert the internal heat of refrigeration into mechanical energy, the compressed gas can be expanded through an orifice and its temperature lowered by the Joule– Thompson effect, or energy can be recovered by extracting work from the expanding gas in an engine.
To liquefy a boiling gas, it has to be cooled below its dew point (the temperature at which condensation starts taking place at a given pressure). The table above lists dew points for a number of gas pressures. To cool a gas, heat energy has to be removed from the compressed gas either by means of cooling water, if this is possible, or by means of an evaporating refrigerant if the temperature of heat removal is lower than ambient. Table shows that only propane can be liquefied by heat exchange with cooling water at moderately high pressure. All lower boiling gases require refrigeration by means of a refrigerant prior to a final compression step which results in liquefaction.
Since sensible heats of gases are lower than their latent heats, latent heats are used to transfer the bulk of the refrigeration energy. Any intermediate refrigerants will boil at atmospheric or slightly lower pressures, at some intermediate temperature between the boiling point of the gas which is to be liquefied and ambient temperature.
A single refrigerant will as a rule permit cooling by 60–90 °C; if an even lower temperature is required, a number of refrigerants may be needed to act as intermediaries in the overall transfer of heat from the cold gas to cooling water or air.
This cycle uses three separate refrigerants, propane, ethylene and methane, in three compression refrigeration cycles operating at successively lower temperatures with the lower temperature cycles each rejecting heat to the next warmer cycle, i.e. cascading on to it. Operating conditions for a cascade process are therefore largely defined once the number and type of refrigerants have been selected.
Appropriate temperatures and pressures are given to Fig. below which slows a simplified flow sheet for the process. Since cooling water or ambient air is the cheapest means of refrigeration, each gas is first cooled to ambient temperature before being heat exchanged with the condensed refrigerant. Also, each of the refrigerants is cooled by heat exchanger with its own vapour (not shown) before expansion through the throttling valve. Finally, each of the higher boiling refrigerants can be expanded at more than one pressure, e.g. propane liquid at medium pressure pre-cools both ethylene and methane; this is followed by a second expansion at lower pressure to cool methane only. Ethylene is also used at two pressures to cool the methane in two stages before its final expansion (not shown). The first two base load plants built in Algeria and Alaska used cycles of this type.
Modified Cascade Cycles (Mixed Refrigerant Cycles)
In these cycles, refrigerant circulation is confined to a single stream, and by operating a number of stages of heat exchange at different pressures, the mixed refrigerant produced using the heavier hydrocarbons from the natural gas itself can be pumped as a single fluid. Figure below shows a simplified flow diagram for a closed cycle mixed refrigerant cycle plant. The refrigerant stream consists of methane, ethane, propane and butane compressed to about 40 bar in a two stage compressor, cooled with cooling water and split in a knockout drum into a condensed propane/butane stream and an overhead consisting of the lighter components. Both streams, together with the natural gas feed, pass to the first heat exchanger (E.l) at the inlet the heavy stream is expanded to compressor inlet pressure and returned through the third passage of the exchanger. The overhead after pre-cooling in E.l is split in a second knockout vessel, the ethane separator into condensate which is expanded at the inlet of the second exchanger (E.2) through which it passes in counter-current with the light overhead, which also passes through E.2 and goes into E.3. After expansion to compressor inlet pressure, it joins the low-pressure propane/butane vapour from E.l having first exchanged its cold with the liquid ethane and gaseous feed streams in E.2. E.3 serves to further cool the natural gas feed and the uncondensed methane, which is expanded at its outlet to compressor inlet pressure, its flow is reversed, and it is mixed with the other expanded gases after heat exchange in E.3. Liquefaction of the natural gas feed is effected by passing it as a separate medium pressure stream through the three heat exchangers in series and expanding the gas through a throttling valve after the last exchanger.
The main advantages of such modified cycles are the smaller number of compressors and heat exchangers required; they also require less space, are simpler to operate and cost less to build. There is no need for a supply of or storage for refrigerant, and losses are easily made up. By careful selection of operating pressure, refrigerant composition and heat exchanger arrangement, efficient processes can be designed. The first base load plant of this type entered service in Libya in 1970.
Pre-cooled Mixed Refrigerant Cycle
This process was developed later and has been selected for all the most recent LNG plants. It combines the simplicity of a mixed refrigerant cycle with the efficiency of the cascade cycle and is now used at more than one half of the base load LNG plants. The process is now the preferred choice in terms of costs flexibility and efficiency; it consists of a mixed liquefaction refrigerant cycle and a separate cycle for pre-cooling the natural gas feed and the liquefaction refrigerant.
The mixed refrigerant consisting of nitrogen, methane, ethane, propane and possibly some butane is compressed and cooled against an ambient temperature cooling medium and pre-cooling refrigerant. The mixture that is split into a light and heavy stream is liquefied in the bundles of a spool-wound heat exchanger and subsequently expanded into the exchanger shell to serve as refrigerant.
The refrigerant is gradually vaporised and warmed up against natural gas and high-pressure refrigerant in the bundles (auto-cooling). Refrigerant vapour from the bottom of the exchanger shell is recompressed and the loop is closed. A simplified flow diagram is shown in Fig. bellow Normally, pre-cooling of liquefaction refrigerant and natural gas is provided by a propane cycle consisting of kettle-type heat exchangers and three or four pressure stages. The plant built years ago in Nigeria is a double mixed type.
1. Natural Gas Engineering ans Safety Challenges.
2. Natural Gas Engineering Handbook.
3. LNG – Future aspects.