Optimizing NGL treatment for sulfur and methanol reduction at Pembina NGL Corporation’s Redwater fractionation plant—Part 1
Pembina NGL Corp. provides sweet and sour gas gathering, compression, condensate stabilization, and both shallow-cut and deep-cut gas processing services, with a total capacity of approximately 6 Bft3d. Nearly all the condensate and natural gas liquids (NGL) extracted through Canadian-based facilities are transported by Pembina’s Pipelines Division. In addition, all NGL transported along the Alliance Pipeline are extracted through a Pembina-operated facility at the pipeline’s terminus in Channahon, Illinois.
Pembina’s Facilities Division includes approximately 354,000 bpd of NGL fractionation, 21 MMbbl of cavern storage, and associated pipeline and rail terminal facilities. The company also recently completed the construction of a liquefied propane export facility on Canada’s west coast. These facilities are fully integrated with Pembina’s other divisions, providing customers the ability to access a comprehensive suite of services to enhance the value of their hydrocarbons. In addition, Pembina owns a bulk marine export terminal in Vancouver, British Columbia.
Pembina also operates a plant site in Redwater, Alberta, Canada. This plant receives raw NGL from field pipelines that is then fractionated into ethane, propane, butane and C5+ condensate. The final products are transported by pipeline, rail and truck, depending on the commodity.
The main contaminants that Pembina is concerned with are sulfur (mercaptans), water and methanol in the final products of butane and propane. Mercaptans and water are naturally occurring compounds, while the methanol is a result of intentional upstream injections to reduce the formation of hydrates in pipelines. Currently, there are three fractionation trains designated as RFS1, RFS2 and RFS3, with a fourth train (RFS4) under construction.
A map1 of the Pembina facilities, including the Redwater complex, is shown in FIG. 1. The contaminant specifications for propane and butane are given in TABLE 1.
FIG. 1. A map of Pembina’s facilities.
In addition to the need to meet final product purity specifications, the reduction of methanol was found to be beneficial to extend the cycles between regeneration in the dryer systems that included both salt and molecular sieve adsorbents, depending on the different plant trains. Methanol will occupy adsorbent sites that water would normally fill, thereby consuming dryer bed capacity.
A licensed extractive Merox processa,3 is used to remove mercaptans from the propane and butane streams with an alkaline solution. The process oxidizes the resultant sodium mercaptides into disulfide oil that is then separated out of the regenerated caustic. To conserve capital spending (CAPEX), the propane and butane streams share a common caustic treating and oxidizing Merox unit. A process schematic showing the process is provided in FIG. 2.
FIG. 2. The Pembina process with fractionation and Merox treating, showing two methanol-water extraction methods.
The process illustration in FIG. 2 shows the two water extraction methods evaluated on the butane stream. Similar water extraction is also used for the propane stream. The two water extraction methods consist of either the use of a static mixer with a vertical mesh pad drum or the use of a mix valve with a high-efficiency cartridge pre-filter/liquid-liquid coalescer unit. Depending on the train involved (RFS1, RFS2 or RFS3), either a molecular sieve bed or a salt dryer is used, or both, to remove the remaining water to get to below the saturation level. When methanol is present, it consumes adsorption sites on the molecular sieve or salt that would otherwise be used for water, which can risk an off-spec event if not controlled by other means. The inefficient separation of water/caustic drops in the form of emulsions caused by the methanol entering the extraction process can also affect dryer bed performance.
Methanol was found to concentrate in the propane leaving the fractionator, which is then extracted by the Merox extractor, where it accumulates in the caustic stream. Since the propane and butane are using a common caustic regeneration section, the methanol can cross over to the butane stream.
Butane segregation test. To verify that the common caustic stream was transferring methanol from the propane to the butane, a series of plant trials was conducted. On both Train RFS2 and Train RFS3, the propane contactor was isolated from the caustic stream by adding a bypass section of piping, closing the valves on the lean caustic entering the propane contactor and closing the valve on the rich caustic leaving the propane contactor. With this modified process in place, the caustic was only treating the butane and was not exposed to the propane circulating caustic or caustic addition/transfer piping.
The bypass process modification is shown in FIG. 3.
FIG. 3. Caustic bypass schematic.
The methanol concentration in the caustic was measured over time, and, once the methanol source was isolated (incoming propane), the methanol levels began to decrease and eventually reached zero concentration. This data is presented in FIG. 4.
FIG. 4. Methanol reduction in butane with the isolated caustic stream.
Based on FIG. 4, the methanol concentration in the butane drops down to zero over a few days once the recirculating caustic stream is isolated. Clearly, the source of the methanol is the propane stream, and the inlet butane contains no methanol. While this bypass modification was useful in confirming the methanol source, it was not practical to operate this way for the long term, as it reduced the system capacity to remove mercaptans to meet the sulfur specifications and left the propane with only the pre-caustic treatment step, which would be insufficient for any spikes in inlet sulfur contaminants.
In April 2020, Pembina modified the routing of the caustic streams from the propane extractors of Trains RFS2 and RFS3 with additional piping to completely segregate them so that the caustic from both extractors would be treated in the RFS3 Merox regeneration section—with the same caustic and Merox system. A similar arrangement was made for the butane extractor units, sending their rich caustic to the RFS2 Merox regeneration section.
Waterwash optimization. The butane waterwash system initially consisted of a recirculating water stream that was injected into the butane and then passed through an inline static mixer before going to a knockout (KO) drum with a mesh pad coalescer. Fresh water was added to the recirculating water stream at about 1 vol% of the butane flowrate, and spent water containing methanol was removed at the same rate. This system is depicted at the top right of FIG. 2.
Several factors affect how the waterwash system performs, including the:
- Flowrate of freshwater injection/spent water removal
- Water recirculation rate
- Mixing method (shear)
- Water level in the separation KO drum
- Separation technology—KO drum with mesh pad or high-efficiency coalescer cartridges.
Waterwash modifications. After reviewing the operation of the waterwash systems in RFS2 and RFS3, several changes were made to improve the overall methanol removal efficiency:
- Pressure drop across the static mixer: The differential pressure across the static mixer affects the shear given to the two-phase system. If the pressure drop is too low, there will be insufficient mixing and poor mass transfer of the methanol from the butane to the aqueous phase. If the pressure drop across the static mixer is too high, this can result in the formation of emulsions with very small drops—in the single-digit micron-size range—that will not be separated by the KO drum with a mesh pad separator.
- Recycle rate of water: The recycle water rate affects the pressure drop across the static mixer: lowering this rate was found to be beneficial in reducing the shear and avoiding the formation of emulsions that could reduce the separation efficiency of the KO drum with a mesh pad.
- Water level in the KO drum: Water and butane are mixed across the static mixer after they enter the KO drum at the bottom section and rise to the top where the mesh pad is located. The butane then exits the vessel, while the water stays in the vessel. By increasing the interface level of water in the KO drum, the path that the butane travels in the water phase is increased, and, therefore, the mass transfer of the methanol from the butane to the aqueous phase is improved.
- Freshwater addition rate: Increasing the freshwater addition will improve the methanol removal efficiency; however, it will also create more wastewater that must be processed. For the Pembina waterwash systems, typical water addition rates were 1%–1.5% of the butane or 1.5%–2% of propane flowrates.
The field results before and after these changes are presented in TABLE 2 and TABLE 3 for Trains RFS2 and RFS3, respectively.
The RFS2 train’s waterwash system was modified by lowering the static mixer differential pressure from an average of 47 kPa down to 20 kPa. The water recirculation rate was lowered from an average of 19 m3/hr to 12 m3/hr, and the KO drum water level was increased from an average of 40% to 66%. These changes resulted in the methanol average removal performance increasing from 64% to 83%, lowering the methanol concentration average in the outlet butane from 62 ppm to 28 ppm. For the RFS2 train, the modifications to the waterwash system reduced the outlet methanol content to consistently meet the 50-ppm specification.
The RFS3 train’s butane waterwash system was modified by taking the static mixer off bypass mode and operating it at an average differential pressure of 22 kPa. The water recirculation rate was lowered from 23 m3/hr to 10 m3/hr, and the KO drum’s water level was increased from an average of 47% to 67%. These changes resulted in the methanol removal performance increasing from 69% to 81%, lowering the methanol concentration in the butane from an average outlet concentration of 105 ppm to 83 ppm. For the RFS3 train, the modifications to the waterwash system were unable to reduce the outlet methanol content to consistently meet the 50-ppm specification.
Part 2. Part 2 will be featured in the May/June issue.
NOTE
a Honeywell UOP
LITERATURE CITED
1 Pembina, “Facilities Division: Natural gas processing and NGL fractionation,” online: https://www.pembina.com/operations/facilities/
2 ASTM D2713-20, “Standard test method for dryness of propane (valve freeze method),” October 7, 2024, online: https://www.astm.org/d2713-20.html
3 Meyers, R. A., “UOP Merox process,” Handbook of Petroleum Refining Processes, 4th Ed., McGraw-Hill Education, 2016.
Comments