3. Dehydration Unit (D-2000)
3.1 Design Concept and Basis
The Dehydration Unit (D-2000) is to remove water from the natural gas feed prior to the pre-cooling of the gas. Three fixed molecular sieve beds operating in parallel are used in the process. At any given time two of the beds are in the adsorption cycle while the other bed is in the regeneration cycle.
Water is adsorbed by the molecular sieves. Each adsorption bed is charged with 30,000 kg of molecular sieves with an adsorptive capacity of 10 kg water/100 kg molecular sieve. On this basis, a maximum of 3,000 kg of water can be adsorbed at the design conditions of column inlet temperature and space velocity. Once the quantity of adsorbed water exceeds this value, water breakthrough will occur, and water will remain in the natural gas stream passing through the dehydrators.
At the design steady state conditions, the treated gas coming from the acid gas removal unit (A-1000) via the feed gas separator (D-V-2003) contains 0.056 mole % of water which means that the water flow rate included in the treated gas is 305 kg/h. As the treated gas is evenly split to two (2) molecular sieve beds which are in adsorption cycle, 152.5 kg/h of water flows to one (1) bed.
Three beds are provided and in normal operation two beds are in parallel adsorption while the third bed is in regeneration and standby. The bed adsorption time of 15 hours has been design based on that one (1) bed adsorbs 2287.5 kg of water which is about 76% of maximum capacity (3,000 kg of water) and hence the proceeding regeneration process is 7½ hours, including a 225 minute cooling and standby time. The sequence control and logic of adsorption and regeneration cycle has to be implemented in DCS.
The treated gas from Unit D-2000 must meet a moisture content specification of <1.0 ppmv maximum.
3.2 Process Theory
In Unit D-2000 water is removed from the feed gas by a process known as physical adsorption. This is a process where the water molecules are held on the surface of a solid desiccant by surface forces. Molecular sieves are used as the adsorbent to remove the water.
Molecular sieves are a crystalline form of a metallo alumino-silicate (zeolite). They can be manufactured with a distinct pore opening in their lattice structure by choosing a particular metal atom. This sizing can make the sieve very selective to which particular molecule it adsorbs.
Type 4A molecular sieves (3.2mm and 1.6mm extrudates or beads) are assumed to be used in the Dehydration Unit. Type 4A (Sodium as the metal atom.) molecular sieves have a pore diameter which is approximately 4 Angstroms (100,000,000A = 1 cm) and will therefore adsorb molecules with a diameter less than approximately 4A.
These include Carbon Dioxide (2.4A), Water (3.2A) and Hydrogen Sulphide (3.6A).
Molecular sieves have electrically active sites and are attracted to dissimilar changes on polar molecules. Water molecules have a strong polarity.
During operation the sieve will adsorb the components listed above. However, the less polar molecules, carbon dioxide and hydrogen sulphide, will be displaced by the water. As the water is adsorbed onto the sieve the other components are displaced and move ahead of the water toward the exit of the bed. Because water is strongly attracted to the molecular sieve it is possible to reduce the water content in the outlet gas to less than 1 ppmv (or a dew point temperature of -90oC or less).
3.3 Process Description
3.3.1 Feed and Adsorption System
The feed to Unit D-2000 is treated gas from the Acid Gas Removal Unit, which is saturated with water. The gas is pre-cooled upstream of Unit D-2000 in the HHP Propane Chiller (L-E-4015) which results in condensation of part of the water vapour contained in the gas. This reduces the load on the Dehydration Unit. The temperature of the feed gas exiting L-E-4015 is maintained at 22oC, a temperature high enough to avoid hydrate formation but low enough to condense a significant amount of water.
The liquid that condenses in L-E-4015 will normally be water that will contain small amounts of heavy hydrocarbons and traces of amine. A significant amount of heavy hydrocarbons will condense in L-E-4015. The condensed liquid is separated from the natural gas in the Feed Gas Separator (D-V-2003) and is sent to the 3 Phase Separator (D-V-2002).
The natural gas from D-V-2003 is sent to the Molecular Sieve Driers (D-C-2001 A/B/C) where the flow is downwards through two parallel driers while the third drier is either regenerating or is on cooldown/standby. As the gas flows through the driers, the water content is reduced to a value below 1 ppmv. Dry natural gas exits the bottom of the two driers and flows through a cartridge filter (D-S-2001) to remove fine dust particles that may have escaped from the driers. The gas then flows to the Mercury Removal Unit (M-3000).
3.3.2 Regeneration System
After a period of time in adsorption mode, the molecular sieve becomes saturated with water and the concentration of water in the gas exiting the driers will begin to increase. Before reaching this condition, the drier is disconnected from the adsorption system and regenerated to remove the water. This restores the molecular sieve to dry conditions suitable for another adsorption cycle. Before a drier is disconnected from adsorption, the standby drier (which has finished regeneration) is connected back to the adsorption system.
The driers are regenerated at high pressure because:
- It reduces the number of ‘pressure cycles’ which the sieve must undergo thereby extending the life of the sieve (pressure cycles can physically damage the sieve particles).
- It reduces the required regeneration heating time. The flow rate of regeneration gas which can be passed up through the bed is limited by the velocity which will ‘lift’ the molecular sieve particles. The velocity limitation exists at both high and low pressure. Because of the higher density at high pressure, a much larger mass flow rate is achieved at the maximum allowable velocity. This reduces the required regeneration heating step time.
The minimum concentration of water in the dried gas, which can be achieved during adsorption, is determined by the degree to which the exit (bottom) of the bed is regenerated or dried. Therefore, the driers are regenerated with reverse flow (entering at the bottom and exiting at the top) because the regeneration gas is driest and hottest when it enters the bed and also that the bottom of the bed is the least saturated area. Reverse flow also avoids the transportation of water from the saturated zone at the top of the bed to the already dry molecular sieve at the bottom, thus improving overall efficiency.
In normal operation, the bed adsorption time is 15 hours and the bed regeneration time is 7.5 hours, including 3.75 hours cooling and standby time. When the bed starts to regeneration, the high temperature dry gas flows from the bottom of the bed for 3.75 hours and after finishing the drying operation, the bed remains in cooling and standby mode for another 3.75 hours. Those adsorption/regeneration cycle sequences shall be implemented in DCS to be operated by automatic logic sequence in normal operation.
To meet the drying cycle above with three molecular sieve beds, when the drying operation starts, the second bed in drying operation starts 7.5 hours after the first bed started so that the first bed can be changed from adsorption operation mode to regeneration mode after 15 hours of drying operation.
Regeneration gas is taken from the dry feed gas line downstream of the driers and the filter (D-S-2001). This gas stream is sent to the Regeneration Gas Heater (D-E-2002) and during heating periods of the regeneration sequence it is heated to 300oC.
Regeneration cycle including heating, cooling and standby modes consists of seven (7) stages.
[table id=37 /]
The hot regeneration gas from D-E-2002 flows upward through the drier bed being regenerated, and absorbs the water contained on the molecular sieves. The regeneration gas is sent to the Regeneration Gas Cooler (D-E-2001), which is an air cooled heat exchanger equipped with cooling fan and cools the regenerated gas to 45°C before the regeneration gas separator (D-V-2001). The regeneration gas stream is cooled during heating periods of the regeneration sequence, which results in condensation of the water vapour contained in the gas. The condensed water which will also contain small amounts of hydrocarbons is separated out in the Regeneration Gas Separator (D-V-2001), and then flows to the 3 Phase Separator (D-V-2002).
The vapour from D-V-2001 is sent to the Regeneration Gas Compressor (D-K-2001). The regeneration gas compressor (D-K-2001) is an electrically driven reciprocating compressor, which raises the pressure of the regeneration gas from 55.25 bara to 62 bara so that it can be recycled upstream of L-E-4015. If the compressor trips during normal operation, the sequence will continue because the regeneration gas is automatically routed to the flare via D-PCV-2001.
The liquid phases from D-V-2001 and D-V-2003 are sent to D-V-2002 where the water and hydrocarbon phases are separated. The water and hydrocarbon are separated in D-V-2002 and sent to the pressure boundaries. Flashed vapour from D-V-2002 is sent to the LP Fuel Gas System (U-6000) through the pressure control valve (D-PCV2002).
Figure 3 shows the overall process of the Dehydration unit (D-1000).
Figure 3 Overall Process of the Dehydration Unit
4. Mercury Removal Unit (M-3000)
Based on the process design principle of the LNG plant and natural gas pre-treatment, the Mercury Removal Unit (M-3000) is installed downstream of the Dehydration Unit (D-2000) to remove trace amount of mercury present in the feed to the Liquefaction Unit (L-4000) in order to protect the aluminium components of the Main Cryogenic Heat Exchanger (L-E-4007) because Mercury, even in trace amount, corrodes aluminium rapidly.
As the design feed gas of this project does not include the mercury, even in trace quantities and the process design does not consider a mechanical corrosion and erosion, the Mercury Removal Unit (M-3000) is designed just to explain the general gas treatment process prior to the liquefaction unit and to consider the pressure drop of the treated dry gas sent to the liquefaction unit.
The dynamic process model shall include the Mercury Removal Unit (M-3000) to consider the pressure drop in the process.
The Mercury Removal Absorber (M-C-3001) can be bypassed by a bypass pipeline and valve installed around the unit M-c-3001. This is to give an operation with no mercury in the feed gas.
Figure 4 shows the overall process of the Mercury Removal Unit (M-3000).
Figure 4 Overall Process of the Mercury Removal Unit
LNG FPSO MPDS (Multi-Purpose Dynamic Simulator) PROJECT SERIES