A facility and a membrane process for cost optimized separating methane and carbon dioxide from a gas stream with very low methane emission

Abstract

The invention is directed at a membrane process and a facility for separating methane and carbon dioxide from a gas stream, providing a methane stream suitable for injection into a natural gas grid, which can achieve low emission of methane to the atmosphere at very low operating and investment costs.

Description

A facility and a membrane process for cost optimized separating methane and carbon dioxide from a gas stream with very low methane emissionField of the invention

[0001] The invention is directed at a membrane process and a facility for separating methane and carbon dioxide from a gas stream, providing a methane stream suitable for injection into a natural gas grid, which can achieve low emission of methane to the atmosphere at very low operating and investment costs.Background of the invention

[0002] Biogas resulting from anaerobic fermentation, such as biogas from an anaerobic digester or a landfill gas, comprises methane and carbon dioxide as the major components. Separating methane from biogas in a quality suitable for feeding the methane into a gas distribution grid is of commercial interest. Membrane processes are advantageous for separating methane from carbon dioxide as they do not require an absorbent for carbon dioxide and can be operated with low energy consumption. Since methane is a more potent greenhouse gas than carbon dioxide, the carbon dioxide enriched stream obtained by a membrane separation process can only be discharged to the atmosphere if it is separated with a low methane content or subjected to an additional treatment for methane removal. Such additional treatment for methane removal consumes energy and requires extra equipment.

[0003] WO 2012 / 000727 discloses a membrane process with three membrane units which can separate biogas into a biomethane stream containing more than 98 vol% methane and a carbon dioxide enriched stream containing about 0.5 % methane at a low recycle rate of less than 60 % which makes the process energy efficient.

[0004] US 7,537,641 B2 discloses a membrane process with four membrane units which is used for treating natural gas produced by gas well that contains an undesirably high concentration of the fracturing gas N2 or CO2. In the process of US 7,537,641 B2 purified methane is obtained in the permeate product stream. The retentate stream which is discharged contains large amounts of methane. Thus, the process disclosed in US 7,537,641 B2 shows insufficient methane recovery, i.e. high methane losses, and can no longer be used under more challenging environmental standards regarding the methane emission.

[0005] Another four-stage membrane separation process is disclosed in EP 2735355 A1 . EP 2735355 A1 discloses a device for separating essentially unpressurized gas mixtures containing methane and carbon dioxide using membrane gas permeation. According to EP 2735355 A1 a maximum methane recovery of 99.5% can be achieved but with very high energy costs. Thus, the process disclosed in EP 2735355 A1 is economically unattractive.

[0006] WO 2015 / 036709 discloses a membrane process with four membrane units which aims at further reducing the energy required for compressing recycled gas but provides a lower methane recovery compared to the process of WO 2012 / 000727. The process provides two carbon dioxide enriched streams from the third and the fourth membrane unit. WO 2015 / 036709 suggests that these two streamsmay be separately or jointly treated by thermal oxidation, used for upgrading the carbon dioxide or discharged to the atmosphere.

[0007] Finally WO 2022 / 012944 A1 discloses a membrane process with four membrane units which aims at further optimizing the process disclosed in WO 2015 / 036709 A1 to enable use at locations with high environmental standards with regard to methane emission.

[0008] On September 24, 2018 the Oil and Gas Climate Initiative (OGCI) published a first methane emission target for its member companies. A base line for methane that gets lost when producing oil and gas of max. 0.32 % and a target of 0.25% methane loss for 2025 was set.

[0009] Tightened regulations on emission of greenhouse gases, e.g. §36 of the German “42. Verordnung uber den Zugang zu Gasversorgungsnetzen (Gasnetzzugangsverordnung - GasNZV)“, require even more ambitious targets for lowering methane emissions from biogas upgrading or natural gas purification (max. 0.2 %). The prior art membrane processes can achieve such goals only by significantly high recycle rates or by an additional step of removing methane from the carbon dioxide enriched streams before discharge to the atmosphere. Both measures increase costs and decrease efficiency of the prior art processes.

[0010] Therefore, a strong need remains for an efficient process for separating methane and carbon dioxide from a gas stream, which fulfills the requirement of the tightened regulations on emissions of greenhouse gases with little extra equipment and energy consumption.

[0011] Subject of the present invention was to provide a new facility and a new process having the disadvantages of the prior art processes and facilities to a reduced degree respectively not having the disadvantages of the prior art processes and facilities.

[0012] A specific problem of the present invention was to provide a new facility and a new process for separating methane and carbon dioxide from a gas stream, which fulfills the requirements of tightened regulations on emissions of greenhouse gases, in particular with regard to gas streams that are discharged to the atmosphere and that should have a methane content of below or equal to 0.5 vol%, preferably below or equal to 0.3 vol% and particularly preferably less than 0.2 vol%.

[0013] Another specific problem of the present invention was to provide a new facility and a new process for separating methane and carbon dioxide from a gas stream, wherein at least one carbon dioxide enriched stream, that is discharged to the atmosphere, is provided having a methane content of below or equal to 1 .5 vol%, preferably below or equal to 1 .3 vol% , more preferred below or equal to 1 .0 vol%, more preferred below or equal to 0.5 vol%, particularly preferably less than 0.3 vol% and most preferred below or equal to 0.2 vol%, preferably without oxidative, methane removing post treatment step.

[0014] In another specific problem of the present invention a new facility and a new process for upgrading a gas comprising methane and carbon dioxide shall be provided, wherein a methane product stream having a methane content of more than or equal to 97 vol% can be obtained and simultaneously a high methane yield, i.e. methane recovery of more than 99%, preferably more than 99.5%, more preferred 99.7% can be achieved.

[0015] In another specific problem of the present invention a new facility and a new process for upgrading a gas comprising methane and carbon dioxide shall be provided, that fulfills the above definedrequirements regarding methane purity of the methane product stream, methane recovery and methane content in the CO2 product stream, which are highly efficient in view of operating costs and / or invest costs, in particular having the lowest possible energy consumption.

[0016] In another specific problem of the present invention a new facility and a new process for upgrading a gas comprising methane and carbon dioxide shall be provided, allowing to continuously fulfill regulatory requirements with regard to methane emission to the atmosphere even if the composition and / or flow rate of the raw gas stream vary.

[0017] In another specific problem of the present invention a new facility and a new process for upgrading a gas comprising methane and carbon dioxide shall be provided, that provides a methane enriched product stream at a pressure that is as significantly above atmospheric pressure, so that it can be fed into a pressurized gas network, for example into CBG (compressed biogas) or LBG (liquefied biogas), or can be further processed with comparatively lower energy expenditure. The new facility and a new process should preferably allow to flexibly adapt the methane purity and the pressure of the methane enriched product gas stream.

[0018] Further problems solved by the present invention but not described before, can be derived from the subsequent description, examples, figures and claims.Detailed Description

[0019] Before describing the invention in more details, some important terms are defined as follows:

[0020] The verb “to comprise” as is used in the description, examples and the claims and its conjugation is used in its non-limiting sense to mean that items following the word are included, but items not specifically mentioned are not excluded. “Comprising” includes “consisting of’ meaning that items following the word “comprising”, are included without any additional, not specifically mentioned items, as preferred embodiment.

[0021] Reference to an element be the indefinite article “a” or “an” does not exclude the possibility that more than one of the elements is present, unless the context clearly requires that there is one and only one of the elements. The indefinite article “a” or “an” thus usually means “one or more”.

[0022] All selectivities referred to in the specification, examples and claims are mixed gas selectivities under standard conditions. “Standard conditions” means temperature of 25°C, feed pressure of 9 barg, composition of the gas mixture tested equal to 60% CH4 and 40% CO2, stage cut equal to 30%, pressure on the permeate side of the membrane of 0,05 barg.

[0023] The inventor of the present invention has now surprisingly found that the problems described above, can be solved by using a membrane separation facility with four membrane units as known from US 7,537,641 B2, Figure 2, which facility has been modified by a. passing a fourth retentate stream (18a) through a pressure control unit (22), preferably a pressure reduction valve (22), to obtain a fourth retentate stream (18b) at the pressure of the gas stream (1)plus minus 10%, and recirculating the fourth retentate stream (18b), to the gas stream (1) at a first gas recycle point (19) located up-stream of the main compressor (2), b. configuring and operating the main compressor (2) and the support compressor (9) such that the main compressor (2) has a higher compressor load than the support compressor (9), such that ratio of the power consumption of the support compressor (9) to the power consumption of the main compressor (2) CpOwer(9) / CpOwer (2) is in a range of from 0.001 to 0.8, preferably 0.005 to 0.6, more preferred 0.01 to 0.5, even more preferred 0.01 to 0.4, particularly preferred 0.01 to 0.3 and most preferred 0.01 to 0.25. In US 7,537,641 B2 the main compressor (2) has a lower compressor load than the support compressor (9) (see Tables 4 and 8), c. configuring and operating the support compressor (9) such that the ratio of the outlet pressure of the support compressor (9) to the inlet pressure of the support compressor (9) psupport-out / psupport-in is in a range of from 1 .05 to 3.5, preferably 1.1 to 3, more preferred 1 .1 to 2.5 and most preferred 1 .1 to 2.2. In US 7,537,641 B2 the ratio of the outlet pressure to the inlet pressure of the support compressor (9) is more than 3.6. d. using membranes with a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 25 to 100 in separation stages 1 , 2 and 3 and of 20 to 100 in separation stage 4. As consequence, a methane enriched product gas stream is obtained as second retentate stream rather than in the third permeate stream as in US 7,537,641 B1. In US 7,537,641 B2 membranes with a SCO2 / CH4 = 10 to 15 are used.

[0024] The facility and the process of the invention allow for adhering to strict regulatory requirements for methane emission to the atmosphere for the third permeate stream, even if the third permeate stream is not subjected to methane removing post treatment and directly discharged to the atmosphere. As shown in the examples below, the process of the invention allows to achieve a methane emission via the third permeate stream of below 0.5 vol% or even below 0.3 vol% or below 0.2 vol%. without oxidative post treatment. US 7,537,641 B1 discloses a maximum methane recovery of 95% in Example 2.

[0025] The achievement to provide a third permeate stream with a methane content of 0.3 vol% or below after the membrane separation allows to reduce invest costs since no equipment for oxidative methane removal in the facility and process of the invention is needed.

[0026] Also, the operating costs for methane removal could be reduced compared to the prior art as shown in the examples and comparative examples below. In preferred embodiments of the invention, it was achieved to minimize the total power consumption needed for recompression of the gas streams, which minimizes operational costs as well as invest costs since smaller compressors can be used.

[0027] The examples below further show that by simultaneous minimization of the total power consumption needed for recompression of the gas streams, the process and device of the invention allows to achieve the high methane recovery described before as well as to obtain the two high purity product streams at minimal separation capacity of the membranes used, i.e. lower separation capacity compared to the prior art, which in addition lowers the investment costs.

[0028] Compared to prior art processes the facility and process of the invention, thus, can be operated with minimum costs for recompression and investment even though tightened requirements for methane emission to the atmosphere are fulfilled.

[0029] Process and facility of the invention provide methane product stream having very high methane contents and very high methane yield.

[0030] Process and facility of the invention are flexible with regard to the methane and CO2 contents of the raw gas streams, the required product purity of the product streams, and the pressure of the methane enriched product gas stream, so that it can be fed into a pressurized gas network, for example into CBG (compressed biogas) or LBG (liquefied biogas), as shown in the examples below.

[0031] Further advantages of the facility and the process of the invention are revealed in the description, examples, figures and claims.

[0032] Subject of the invention is therefore a process for separating methane from a gas stream comprising methane and carbon dioxide, comprising the following steps:(a) providing a gas-stream (1), comprising a raw gas stream (RGS), comprising methane and carbon dioxide,(b) compressing the gas stream (1) with a main compressor (2) to provide a feed stream (3)(c) passing the feed stream (3) to a membrane separation unit (4) comprising a feed stream membrane separation stage (5), a retentate stream membrane separation stage (6), a permeate stream membrane separation stage (7), a fourth membrane separation stage (8), and a support compressor (9);(d) passing the feed stream (3) to a feed gas inlet of the feed stream membrane separation stage (5) and processing the feed stream (3) in the feed stream membrane separation stage (5) to produce a first permeate stream (10) at lower pressure than the feed stream, which is enriched in carbon dioxide compared to the feed stream (3), and a first retentate stream (11), which is enriched in methane compared to the feed stream (3);(e) passing the first retentate stream (1 1) to the gas inlet of the retentate stream membrane separation stage (6) and processing the first retentate stream (11) in the retentate stream membrane separation stage (6) to produce a second permeate stream (12) at lower pressure than the feed stream (3), which is enriched in carbon dioxide compared to the first retentate stream (11), and a second retentate stream (13), which is enriched in methane compared to the first retentate stream (1 1);(f) withdrawing the second retentate stream (13) from the membrane separation unit (4) as first product stream or further processing the second retentate stream (13), preferably outside of the membrane separation unit (4);(g) recompressing the first permeate stream (10) with the support compressor (9) to obtain a recompressed first permeate stream (14);(h) passing the recompressed first permeate stream (14) to the gas inlet of the permeate stream membrane separation stage (7) and processing the recompressed first permeate stream (14) in the permeate stream membrane separation stage (7) to produce a third permeate stream (15) at lower pressure than the recompressed first permeate stream (14), which is enriched in carbondioxide compared to the recompressed first permeate stream (14), and a third retentate stream (16), which is enriched in methane compared to the recompressed first permeate stream (14);(i) withdrawing the third permeate stream (15) from the membrane separation unit (4) as second product stream or further processing the third permeate stream (15), preferably outside of the membrane separation unit (4), or sending the third permeate stream (15) to vent;(j) passing the third retentate stream (16) to the gas inlet of the fourth membrane separation stage (8) and processing the third retentate stream (16) in the fourth membrane separation stage (8) to produce a fourth permeate stream (17) at lower pressure than the third retentate stream (16), which is enriched in carbon dioxide compared to the third retentate stream (16), and a fourth retentate stream (18a), enriched in methane compared to the third retentate stream (16);(l) recirculating the fourth permeate stream (17) to the first permeate stream (10), at a third gas recycle point (21) located up-stream of the support compressor (9);(m) recirculating the second permeate stream (12) to the gas stream (1) at a second gas recycle point (20) located up-stream of the main compressor (2) or to the fourth retentate stream (18b) at a fourth gas recycle point located up-stream of the connecting point (19) and downstream of the pressure control unit (22); characterized in that it further comprises the step(k) passing the fourth retentate stream (18a) through a pressure control unit (22), preferably a pressure reduction valve (22), to obtain a fourth retentate stream (18b) at the pressure of the gas stream (1) plus minus 10%, and recirculating the fourth retentate stream (18b), to the gas stream(1) at a first gas recycle point (19) located up-stream of the main compressor (2);(n) optionally but preferably controlling the pressure on the retentate side of the retentate stream membrane separation stage (6), preferably also of the feed stream membrane separation stage (5), with a pressure control unit (23) located in the second retentate stream (13);(o) optionally but preferably controlling the pressure on the retentate side of the fourth membrane separation stage (8), preferably also on the retentate side of the permeate stream membrane separation stage (7), with a pressure control unit (22) located in the fourth retentate stream (18); and that the main compressor (2) and the support compressor (9) are selected and operated such that the main compressor (2) has a higher compressor load than the support compressor (9), such that ratio of the power consumption of the support compressor (9) to the power consumption of the main compressor (2) CpOwer(9) / CpOwer (2) is in a range of from 0.001 to 0.8; the ratio of the outlet pressure of the support compressor (9) to the inlet pressure of the support compressor (9) psupport-out / psupport-in is in a range of from 1 .05 to 3.5; the ratio of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) Pmain-out / Psupport-out * 1 ; the membranes used in separation stages 1 , 2 and 3 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 25 to 100; the membranes used in separation stage 4 having a mixed gas selectivity Sco2 / cH4 at standard pressure and standard temperature of the respective membrane separation unit of 20 to 100.

[0033] The process of the invention is characterized in that a small and cheap support compressor (9) compared to the main compressor (2) is used, i.e. that the main compressor (2) has a higher compressor load than the support compressor (9). Compared to a concept wherein the main compressor (2) has a lower compressor load than the support compressor (9), as disclosed in US 7,537,641 B2, the inventive configuration leads to lower energy consumption, i.e. operating costs, as shown in Examples 3 to 8, if the same raw gas is used and the same specification for the product gases is applied. It was surprising, that in the configuration of the invention a higher gas volume that needs to be compressed with the main compressor (2), can be overcompensated by the reduction of gas volume to be compressed with the support compressor (9). This was particular surprising, since two recycling streams were recycled back to the gas stream (1) in the present invention. Thus, the main compressor (2) and the support compressor (9) are configured such, that the ratio of the power consumption of the support compressor (9) to the power consumption of the main compressor (2) CpOwer(9) / CpOwer (2) is in a range of from 0.001 to 0.8, preferably 0.005 to 0.6, more preferred 0.01 to 0.5, even more preferred 0.01 to 0.4, particularly preferred 0.01 to 0.3 and most preferred 0.01 to 0.25.

[0034] Inventors surprisingly found out that small pressure increase of the first permeate stream (10) to obtain the recompressed first permeate stream (14) is sufficient to solve all problems of the invention, in particular, to obtain the two high purity product streams as well as to achieve a very high methane recovery. In contrast to US 7,537,641 B2, where the pressure increase generated by the compressor in the 1stpermeate stream is minimum 3.6 times of the pressure of the uncompressed feed stream, the pressure increase generated by the support compressor (9) in the present invention, i.e. the ratio of the outlet pressure of the support compressor (9) to the inlet pressure of the support compressor (9) psupport-out I psupport-in is in a range of from 1 .05 to 3.5, preferably 1.1 to 3, more preferred 1 .1 to 2.5 and most preferred 1 .1 to 2.2. This allows to significantly reduce the energy needed to operate the support compressor (9) compared to for example the compressor in the 1stpermeate stream of US 7,537,641 B2.

[0035] The inventors further found out that the energy consumption of the process and facility of the invention can be further reduced if the pressure of the recompressed first permeate stream (14) is lower than the pressure of the feed stream (3). This allows to operate the support compressor (9) with minimal energy consumption. Thus, the ratio of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) Pmain-out I Psupport-out > 1 , preferably is in a range from 1 .1 to 40, more preferred 1 .5 to 16, even more preferred 1 .8 to 12, especially preferred 2 to 10 and most preferred 3 to 6. The inventive process and device configuration, thus, can be operated at lower total energy consumption compared to the process and apparatus disclosed in EP2735355A1 , Figure 5. This is shown in Comparative Example 1 , below.

[0036] Examples 3 to 8 below show that use of higher selective membranes in separation stages 1 to 3 allows to separate an identical raw gas to an identical product gas at lower operational costs. The use of higher selective membranes contribute to a reduction of the total power consumption required by the two compressors in the inventive process and facility. Thus, the membranes used in separation stages 1 , 2 and 3 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 25 to 100, preferably 30 to 100, more preferred 30 to 90, even more preferred 35 to 90, particular preferred 40 to 80 and most preferred 45 to 70. The membranes usedin separation stages 1 , 2 and 3 may have identical or different mixed gas selectivities SCO2 / CH4. Preferably the selectivity ratio of stage 1 to stage 2 is in a range of from 1 to 3, preferably 1 to 2, most preferred 1.1 to 2 and / or the selectivity ratio of stage 1 to stage 3 is in a range of from 0.7 to 1 .2, preferably 0.9 to 1 .1 , most preferred 0.98 to 1 .02. Especially preferred the membranes used in stages 1 and 3 have the same selectivity.

[0037] Example 1 , compared to Example 2 shows that it is possible to use in the fourth separation stage a membrane with lower mixed gas selectivity SCO2 / CH4 than in separation stages 1 to 3 but still to solve the problems of achieving two high purity product streams and achieving the required, very high methane recovery. In this case, the total power consumption increases only slightly compared to the case where identical, high selective membranes are used in all four stages. Thus, the membranes used in separation stage 4 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 20 to 100, preferably 25 to 90, more preferred 25 to 80 and most preferred 25 to 70. More preferred the membranes used in separation stage 4 having an equal or lower, more preferred an equal, mixed gas selectivity SCO2 / CH4, compared to the membranes used in separation stage 1 and / or separation stage 2 and / or separation stage 3. Even more preferred the selectivity SCO2 / CH4 ratio of stage 2 to stage 4 is in a range of from 0.7 to 3, preferably 0.9 to 2, most preferred 1 to 1 .8. Most preferred the membranes used in separation stage 4 having the lowest mixed gas selectivity SCO2 / CH4 of all stages.

[0038] Suitable membranes which have higher permeability for carbon dioxide than for methane are known from the prior art. In general, membranes containing a separation layer of a glassy polymer, i.e. a polymer having a glass transition point at a temperature above the operating temperature of the membrane separation stage, will provide higher permeability for carbon dioxide than for methane. The glassy polymer may be a polyetherimide, a polycarbonate, a polyamide, a polybenzoxazole, a polybenzimidazole, a polysulfone or a polyimide. The gas separation membrane preferably comprises at least 80 % by weight of a polyimide or a mixture of polyimides.

[0039] In a preferred embodiment, the gas separation membrane comprises at least 50 % by weight of a polyimide prepared by reacting a dianhydride selected from 3,4,3’,4’-benzophenonetetracarboxylic dianhydride, 1 ,2,4,5-benzenetetracarboxylic dianhydride, 3,4,3’,4’-biphenyltetracarboxylic dianhydride, oxydiphthalic dianhydride, sulphonyldiphthalic dianhydride, 1 ,1 ,1 ,3,3,3-hexafluoro- 2,2-propylidenediphthalic dianhydride and mixtures thereof with a diisocyanate selected from 2,4-tolylene diisocyanate, 2,6-tolylene diisocyanate, 4,4’-methylenediphenyl diisocyanate, 2,4,6-trimethyl-1 ,3- phenylene diisocyanate, 2,3,5,6-tetramethyl-1 ,4-phenylene diisocyanate and mixtures thereof. The dianhydride is preferably 3,4,3’,4’-benzophenonetetracarboxylic dianhydride or a mixture of 3,4,3’,4’-benzophenonetetracarboxylic dianhydride and 1 ,2,4,5-benzenetetracarboxylic dianhydride. The diisocyanate is preferably a mixture of 2,4-tolylene diisocyanate and 2,6-tolylene diisocyanate or a mixture of 2,4-tolylene diisocyanate, 2,6-tolylene diisocyanate and 4,4’-methylenediphenyl diisocyanate. Suitable polyimides of this type are commercially available from Evonik Fibres GmbH under the trade name P84® type 70, which has CAS number 9046-51-9 and is a polyimide prepared from 3, 4,3’, 4’- benzophenonetetracarboxylic dianhydride and a mixture of 64 mol% 2,4-tolylene diisocyanate, 16 mol% 2,6-tolylene diisocyanate and 20 mol% 4,4’-methylenediphenyl diisocyanate, and under the trade name P84® HT, which has CAS number 134119-41-8 and is a polyimide prepared from a mixture of 60 mol%3, 4, 3’, 4’-benzophenonetetracarboxylic dianhydride and 40 mol% 1 ,2,4,5-benzenetetracarboxylic dianhydride and a mixture of 80 mol% 2,4-tolylene diisocyanate and 20 mol% 2,6-tolylene diisocyanate. The gas separation membranes of this embodiment have preferably been heat treated in an inert atmosphere as described in WO 2014 / 202324 A1 to improve their long-term stability in the process of the invention.

[0040] In another preferred embodiment, the gas separation membrane comprises at least 50 % by weight of a block copolyimide as described in WO 2015 / 091122 on page 6, line 20 to page 16, line 4. The block copolyimide preferably comprises at least 90 % by weight of polyimide blocks having a block length of from 5 to 1000, preferably from 5 to 200.

[0041] The gas separation membrane may be flat membrane or a hollow fiber membrane and is preferably an asymmetrical hollow fiber membrane comprising a dense polyimide layer on a porous support. The term “dense layer” here refers to a layer which comprises essentially no macropores extending through the layer and the term “porous support” here refers to a support material having macropores extending through the support. The asymmetrical hollow fiber membrane can be prepared by coating a porous hollow fiber with a polyimide to form a dense polyimide layer on the support. In a preferred embodiment, the asymmetrical hollow fiber membrane is a membrane prepared in a phase inversion process by spinning with an annular two component spinning nozzle, passing a solution of a polyimide through the annular opening and a liquid containing a non-solvent for the polyimide through the central opening.

[0042] The gas separation membrane preferably comprises a dense separation layer of a glassy polymer coated with a dense layer of a rubbery polymer which rubbery polymer has higher gas permeability than the glassy polymer. The preferred gas separation membranes comprising a polyimide separation layer are preferably coated with a polydimethylsiloxane elastomer.

[0043] When the gas separation membrane is a flat membrane, the membrane separation units preferably comprise one or several spiral wound membrane modules containing the flat membranes and when the gas separation membrane is a hollow fiber membrane the membrane separation units preferably comprise one or several membrane modules containing a bundle of hollow fiber membranes. Each of the membrane separation units may comprise several membrane modules arranged in parallel and may also comprise several membrane modules arranged in series, wherein in a series of membrane modules the retentate provided by a membrane module is passed as feed to the membrane module subsequent in the series of membrane modules, the last membrane module of the series providing the retentate of the membrane separation stage, and the permeates of all membrane modules within a series are combined to provide the permeate of the membrane separation unit. When a membrane separation unit comprises several membrane modules arranged in series, the membrane modules are preferably removable membrane cartridges arranged in series as a chain of cartridges in a common pressure vessel and connected to each other by a central permeate collecting tube, as described in detail inWO 2016 / 198450 A1. Membrane separation units which comprise several membrane modules arranged in parallel are preferred.

[0044] If it is desired to inject the methane enriched product gas stream into a pressurized gas network, for example into CBG (compressed biogas) or LBG (liquefied biogas) without the use of an additionalcompressor arranged downstream of the retentate gas outlet of the retentate stream membrane separation stage (6) it is preferred to adjust the pressure pmain-out at the gas outlet of the main compressor (2) to a range of from 6 to 40 bara, preferably 6 to 30 bara, more preferred 6 to 20 bara, even more preferred 8 to 16 bara and most preferred 8 to 12 bara.

[0045] It is further preferred to arrange a pressure control unit (23) in or downstream to the second retentate stream (13). Pressure control unit (23) allows to control the pressure on the retentate sides of the feed stream membrane separation stage (5) as well as of the retentate stream membrane separation stage (6). If the pressure of the retentate side of the feed stream membrane separation stage (5) shall be adjusted independently from the pressure on retentate side of the retentate stream membrane separation stage (6) another pressure control unit (26) can be arranged in the first retentate stream (11). Use of the pressure control units (23) and (26) allows to flexibly adjust the pressures in the separation stages (5) and (6) as well as the driving forces over said separation stages.

[0046] In the inventive process and apparatus, the fourth retentate stream (18a) is passed through a pressure control unit (22), to obtain a fourth retentate stream (18b) at the pressure of the gas stream (1) plus minus 10%, preferably plus minus 5%, more preferred plus minus 2%, even more preferred plus minus 1 %, most preferred at the pressure of the gas stream (1) and recirculating the fourth retentate stream (18b). The fourth retentate stream (18b) needs to have similar pressure compared to the gas stream (1) to combine both streams to a combined gas stream (1). It is, thus preferred that the ratio of the pressure of the gas-stream (1) to the pressure of the fourth retentate stream (18b) pgsi / pret4 0.98 to 1 .02 most preferred is 1 .

[0047] Pressure control unit (22) allows to control the pressure on the retentate sides of the permeate stream membrane separation stage (7) as well as of the fourth membrane separation stage (8). If the pressure of the retentate side of the permeate stream membrane separation stage (7) shall be adjusted independently from the pressure on retentate side of the fourth membrane separation stage (8) another pressure control unit (27) can be arranged in the third retentate stream (16). Use of the pressure control units (22) and (27) allows to flexibly adjust the pressures in the separation stages (7) and (8) as well as the driving forces over said separation stages.

[0048] The second permeate stream (12) is recycled back to the gas stream (1) at a second gas recycle point (20) located up-stream of the main compressor (2) or to the fourth retentate stream (18b) at a fourth gas recycle point located up-stream of the connecting point (19) and downstream of the pressure control unit (22). Thus, it is preferred that the ratio of the pressure of gas-stream (1) to the pressure of the second permeate stream (12) pgsi / pperm2 is in the range of 0.98 to 1 .02 and most preferred is 1 or that the pressure of fourth retentate stream (18b) to the pressure of the second permeate stream (12) pret4 / pperm2 is in the 0.98 to 1 .02 and most preferred is 1 . The pressure of the second permeate stream (12) can be adjusted in different ways. In a first preferred embodiment, the pressure of the second permeate stream (12) at the permeate gas outlet of the retentate stream membrane separation stage (6) is controlled via the control of the flow amount through compressor (2) and / or the pressure control unit (23) in the second retentate stream (13). In a second preferred embodiment a pressure control unit (24) is arranged in the second permeate stream (12). Depending on the target pressure of the second permeate stream (12),pressure control unit (24) is preferably a control valve or a vacuum blower. In this case a second permeate stream (12a) is generated at the permeate gas outlet of the retentate stream membrane separation stage (6) that is passed through the pressure control unit (24) to obtain a second permeate stream (12b) at the downstream side of the pressure control unit (24). The second permeate stream (12b) having the target pressure defined before for the second permeate stream (12). Depending on the type of pressure control unit (24) the pressure in the second permeate stream (12b) may be higher or lower than the pressure in stream (12a). Use of a pressure control unit (24) allows to flexibly adjust the pressure of the second permeate stream (12a) and thus the stage cut of the retentate stream membrane separation stage (6). It is thus preferred to use such pressure control unit (24) in the process and device of the present invention. The measures of the first and second preferred embodiments can be combined to control the pressure of the second permeate stream (12). The pressure in stream (12a) in the second preferred embodiment respectively stream (12) in the first preferred embodiment is preferably in the range of 0.3 to 1 .5 bara, more preferred 0.5 to 1 .2 bara, even more preferred 0.7 to 1 .2 bara and most preferred 1 .0 to 1 .2 bara.

[0049] In the process and device of the invention the fourth permeate stream (17) is recycled back to the first permeate stream (10) at a third gas recycle point (21) located up-stream of the support compressor (9). It is, thus, preferred that the fourth permeate stream (17) has a similar pressure compared to the permeate stream (10). More preferred the ratio of the pressure of the first permeate stream (10) to the pressure of the fourth permeate stream (17) pPermi / pPerm4 is in the range of 0.9 to 1.1 , preferably 0.95 to 1.05, more preferred 0.98 to 1.02 and most preferred it is 1. The pressure of the fourth permeate stream (17) can be adjusted in different ways. In a first preferred embodiment the pressure of the fourth permeate stream (17) over the fourth membrane separation stage (8) is set by the inlet (suction) pressure of the support compressor (9). In a second preferred embodiment a similar concept as for the fourth retentate stream (18) is applied, i.e. a pressure control unit (25) is arranged in the fourth permeate stream (17). In this case a fourth permeate stream (17a) is generated at the permeate gas outlet of the fourth membrane separation stage (8) that is passed through the pressure control unit (25) to obtain a fourth permeate stream (17b) at the downstream side of the pressure control unit (25). The fourth permeate stream (17b) having the target pressure defined before for the fourth permeate stream (17) and having a lower pressure than the fourth permeate stream (17a). Use of a pressure control unit (25) allows to flexibly adjust the pressure of the fourth permeate stream (17a) and thus of the stage cut SC4 of the fourth membrane separation stage (8). It is thus preferred to use such pressure control unit (25) in the process and device of the present invention.

[0050] Preferably used pressure control units (22), (23) (24), (25), (26) and (27) can be identical or different and are preferably selected, independently from each other from the group of devices consisting of pressure control valves, gate valves, globe valves, ball valves, check valves(spring return valves), butterfly valves, angle valves, ball / plug valves and diaphragm valves.It is further preferred to arrange a first feed stream temperature control unit in the first feed stream (3) and / or a second feed stream temperature control unit in the second retentate stream (11) and / or a third feed stream temperature control unit in the recompressed first permeate stream (14) and / or a fourth feed stream temperature control unit in the third retentate stream (16). Said temperature control units are notshown in the Figures. The temperature control units arranged in the first feed stream (3) and / or the second retentate stream (11) and / or the recompressed first permeate stream (14) and / or the third retentate stream (16) can be identical or different and are preferably selected independently from each other from the group of devices consisting of heat exchangers like tube or plate type heat exchangers, coolers, air blast coolers, fin fan coolers, heater, heater battery, heater coil, heat recovery unit, scrubber, condenser. The membrane separation stages are typically operated at temperatures in a range of from 10 to 50 °C, more preferred 15 to 45 °C, particularly preferred 20 to 40 °C and most preferred 25 to 35 °C. An active adjustment of the temperature difference between the stages is foreseeable but only within a range of 20 K.

[0051] To configure an inventive device and process for a production plant, a man skilled preferably uses a known process simulation software which calculates mass transfer of the gas components through the membrane by numerical integration of the known differential equations for mass transfer through a membrane by a solution-diffusion process. The temperature dependency of permeation can be accounted for by applying the equations known from M. Scholz et. al, Ind. Eng. Chem. Res. 52 (2013) 1079-1088. Such calculations are preferably carried out with boundary conditions set for the target values Cpower(9) / CpOwer (2), pmain-out / Psupport-out mixed gas selectivity SCO2 / CH43raw gas composition, target CF content of the 2ndretentate stream (13), i.e. the first product gas stream, and of the 3rdpermeate stream (15), i.e. the 2ndproduct gas stream, as specified above. Further, methane recovery is preferably used as boundary condition as shown in the examples below. Preferably stage cuts in all four separation stages, pressure ratio over the support compressor (9) psupport-out / psupport-in and permeate side pressure of the first and forth membrane separation (pPermi and pPerm4) are variables and selected from the ranges specified above respectively below by the software as demonstrated in the examples below.

[0052] Inventors found out that low total energy consumption and a low total membrane area requirement can be achieved, if the pressure ppermi of the first permeate stream (10) as well as if the stage cuts of the membrane separation stages (1) to (4) are varied within the boarders defined below:

[0053] The pressure pPermi of the first permeate stream (10) is preferably varied in a range of from 1 bara to 1.2 * jp^feed) bara, preferably of from 1 to [p<jeed) bara, and most preferred from 1 bara to 0.95*jp^feed) bara, wherein p<feed) is the pressure of the feed stream (3).

[0054] The stage cut SC1 of the feed stream membrane separation stage (5) is varied in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 1 mol / mol; more preferred 0.08 to 0.95 mol / mol, particularly preferred 0.15 to 0.65 mol / mol and most preferred 0.18 to 0.62 mol / mol.

[0055] The stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 0.95 mol / mol; more preferred 0.08 to 0.8 mol / mol, particularly preferred 0.09 to 0.75 and most preferred 0.1 to 0.7 mol / mol.

[0056] The stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 0.99 mol / mol; more preferred 0.08 to 0.95 mol / mol and most preferred 0.1 to 0.9 mol / mol.

[0057] The stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 1 mol / mol; more preferred 0.1 to 0.95 mol / mol and most preferred 0.4 to 0.9 mol / mol.

[0058] Preferably stage cuts are varied through variations of one or more parameters selected from the group consisting of composition of the feed stream, operating temperature, feed and permeate pressure, membrane type, e.g. selectivity and / or permeance, and membrane area of the respective separation stage.

[0059] The process and facility of the present invention are preferably operated such that the CO2 concentration in the raw gas stream (RGS) Cco2-raw is controlled to be in a range of from 10 to 70 mol%, preferably 20 to 60 mol% and more preferred 30 to 50 mol% and the methane concentration in the raw gas stream (RGS) CcH4-raw is controlled to be in a range of from 30 to 90 mol%, preferably 40 to 80 mol% and more preferred 50 to 70 mol%, wherein the contents of carbon dioxide and methane are selected such from the ranges defined before that both components plus optionally further gases comprised in the raw gas stream (RGS) sum up to 100 mol% of the raw gas stream (RGS). Inventors found out that less energy for purification is required if Cco2-raw and if CcH4-raw are controlled in the ranges specified before. If the methane concentration in the raw gas stream is too low or if too many other gas components are comprised in the raw gas stream, process efficiency may decrease.

[0060] It is, thus, preferred that the raw gas stream comprises oxygen in a maximum concentration C02- raw of from 0 to 3 mol%, more preferred of from 0.0001 to 2 mol% and most preferred of from 0.001 to 1 mol%.

[0061] It is further preferred that the raw gas stream comprises nitrogen in a maximum concentration CN2-raw of from 0 to 10 mol%, more preferred of from 0.001 to 4 mol% and most preferred of from 0.01 to 2 mol%.

[0062] Preferred raw gases to be separated by the process of the invention are biogas, natural gas or a landfill gas or other gases comprising CO2 and CH4 and optionally O2 and N2 in the above specified ranges. The raw gas is preferably a, more preferred desulfurized, biogas from an anaerobic digester. Desulfurizing the raw gas stream prevents corrosion of the compressor and of gas conduits of the facility. The biogas may also be pretreated by drying and / or by adsorption of volatile organic compounds, such as volatile siloxanes, on an adsorbent to obtain the raw gas. When the raw gas is a biogas from an anaerobic digester operated with controlled air addition to reduce hydrogen sulfide formation in the digester, the raw gas will typically contain minor amounts of oxygen and nitrogen.

[0063] In a preferred embodiment of the present invention the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled by the measures described before, for the inventive process, in particular as boundary condition in the simulation software to be in a range of from 90 to 100 mol%, preferably 95 to 99.9 mol%, more preferred 97 to 99.9 mol%, even more preferred 97 to 99.8 mol%, particularly preferred 97 to 99.5 mol% and most preferred 98 to 99.5 mol%.

[0064] After extensive investigations inventors found out that the process and facility configurations described before can be narrowed down to those having very low ratio of total power consumption to rawflow, i.e. operational costs, and delivering a methane recovery of at least 99.5% if the following additional process parameters are fixed as boundary conditions within the subsequently specified ranges: the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, and the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised in above specified ranges, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bar, and the CO2 / CH4 mixed gas selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 30 to 70, and the ratio pmain-out / psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Vam bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure pPermi of the first permeate stream (10) is in a range of from 1 to Vamti bara, wherein Vamti being the upper limit of the pressure range, and the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Vam being the lower bound and Vars being the upper bound of the stage cut range, and the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from Vam to Vamp mol / mol, wherein Vam being the lower bound and Vamh being the upper bound of the stage cut range, and the stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from Vam to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.1 to 0.9 mol / mol, and wherein Vam, Vanh, Vamti, Varsi, Varsh, Vami, Va h, Varsi and Vanh, are regressed variables calculated by use of equation (I)Var = Interceptwith “Intercept” and “ai” being fixed factors taken from the Table 1 , xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and xi with i = 4 being the selectivity X4 = Sco2 / cH4of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined before for target recovery of at least 99.5%Table 1

[0065] After extensive investigations inventors found out that the process and facility configurations described before can be narrowed down to those having very low ratio of total power consumption to raw flow, i.e. operational costs, and delivering a methane recovery of at least 99.9% if the following additional process parameters are fixed as boundary conditions within the subsequently specified ranges: the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, and the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised can be comprised in above specified ranges, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bar, and the CO2 / CH4 selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 30 to 70, and the ratio Pmain-out / Psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Vam bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure pPermi of the first permeate stream (10) is in a range of from 1 to Vamti bara, wherein Vamti being the upper limit of the pressure range, and the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Vam being the lower bound and Vamh being the upper bound of the stage cut range, and the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from Vam to Vamp mol / mol, wherein Vam being the lower bound and Vamh being the upper bound of the stage cut range, andthe stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from Var5i to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.4 to 0.9 mol / mol if the mixed gas membrane selectivity SCO2 / CH4 is 50 or above 50, preferably 50 to 70, and in a range of from 0.7 to 0.9 mol / mol if the mixed gas membrane selectivity SCO2 / CH4 is 30. For mixed gas membrane selectivity SCO2 / CH4 of 30 to 50 the lower bound of the SC4 range should be calculated scaling linearly with selectivity between the values 0.7 and 0.4 mol / mol (e.g. for selectivity equal to 30 the lower bound is equal to 0.7, for selectivity equal to 40 the lower bound is equal to 0.55 and for selectivity equal to 50 the lower range is equal to 0.4). The upper bound stays at 0.9 mol / mol independent of SCO2 / CH4. wherein Varn, Vanh, Var2h, Varsi, Varsh, Var4i, Var4h, Varsi and Vanh, are regressed variables calculated by use of equation (I) given above with “Intercept” and “ai” being fixed factors taken from the Table 2, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity x4= SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and x4= are values taken from the ranges defined above for target recovery of at least 99.9 %.Table 2

[0066] Inventors further found out that the process and facility configurations described before can be narrowed down to those having the lowest ratio of total power consumption to raw flow, i.e. operational costs, and delivering a methane recovery of at least 99.5% if the following additional process parameters are fixed as boundary conditions within the subsequently specified ranges: the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, andthe CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bara, the pressure (psupport-out) at the gas outlet of the support compressor (9) is controlled to be in a range of from pmain out / Varih to pmain-out / Varn bara, wherein Vam being the lower limit of the range of the pressure ratio pmain-out / psupport-out and Vanh being the upper limit of the pressure ratio range, and the CO2 / CH4 mixed gas selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 30 to 70, and the pressure pPermi of the first permeate stream (10) and correspondingly that of the fourth permeate stream (17) is in a range of from 1 bara to Var2h bara, wherein Var2h being the upper limit of the pressure range, and the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the CO2 concentration in the fourth retentate stream (18) (Cco2(i8>) is controlled to be below an upper bound of Vareh in mole fraction, and the CO2 concentration in the fourth permeate stream (17) (Cco2(i7>) is controlled to be below an upper bound of Var?h in mole fraction, and if Vam, Vanh and Var2h, are regressed variables calculated by use of equation (II)Var = Interceptwith “ai” being the coefficients for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “ai ” and “a^” being fixed factors taken from the Table 3 below, product pairs or squared terms in equation (II) not specified in Table 3 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.5%Table 3and whereinVarsi and Varsh are regressed variables being calculated by use of equation (III)Var = Interceptwith “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 4 below, product pairs or squared terms in equation (III) not specified in Table 4 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.5%Table 4 for Varsi and Varshand wherein Vareh and Var?h are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 5 below, product pairs or squared terms in equation (lll)not specified in Table 5 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.5%Table 5 for Vareh and Var?h

[0067] Inventors further found out that the process and facility configurations described before can be narrowed down to those having the lowest ratio of total power consumption to raw flow, i.e. operational costs, and delivering a methane recovery of at least 99.9% if the following additional process parameters are fixed as boundary conditions within the subsequently specified ranges: the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol% the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too, the pressure (pmain) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bara, the pressure (psupport) at the gas outlet of the support compressor (9) is controlled to be in a range of from pmain-out / Varih to pmain-out / Varn bara, the CO2 / CH4 selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 30 to 70 the pressure pPermi of the first permeate stream (10) is in a range of from 1 bara to Var2h bara,the stage cut SCi of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, the CO2 concentration in the fourth retentate stream (18) (Cco2(i8>) is controlled to be below an upper bound of Vareh in mole fraction, the CO2 concentration in the fourth permeate stream (17) (Cco2(i7>) is controlled to be below an upper bound of Var?h in mole fraction, wherein Varn, Vanh and Var2h, are regressed variables being calculated by use of equation (II) with “ai” being the coefficients for the single variables, “ai ” being interaction coefficients for the product of the given variables, and “ai»2” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 6 below, product pairs or squared terms in equation (II) not specified in Table 6 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi, X2 X3 and X4 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.9%Table 6and wherein Varsi and Varsh, being regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “ai ” and “a^” being fixed factors taken from the Table 7 below, product pairs orsquared terms in equation (III) not specified in Table 7 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi, X2, X3, X4 and X5 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.9%Table 7 for Varsi and Varshand wherein Vareh and Var?h being regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “ai ” and “a^” being fixed factors taken from the Table 8 below, product pairs or squared terms in equation (III) not specified in Table 8 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranesand Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi, X2, X3, X4 and X5 = are values taken from the ranges defined above for the configuration having the lowest operating costs and a methane recovery of at least 99.9%Table 8 for Vareh and Var?h

[0068] A further subject of the invention is a device for separating methane from a gas stream comprising methane and carbon dioxide, comprising: - a gas-stream conduit (1) connected to a raw gas source, providing a raw gas comprising methane and carbon dioxide, and to a gas inlet of a main compressor (2) a feed stream conduit (3) connected to a gas outlet of the main compressor (2) and to a feed gas inlet of a feed stream membrane separation stage (5), which is part of a membrane separation unit (4), which further comprises a retentate stream membrane separation stage (6), a permeate stream membrane separation stage (7), a fourth membrane separation stage (8), and a support compressor (9)a first permeate stream conduit (10) connected to a permeate gas outlet of the feed stream membrane separation stage (5) and to a gas inlet of a support compressor (9) a first retentate stream conduit (11) connected to a retentate gas outlet of the feed stream membrane separation stage (5) and to a gas inlet of the retentate stream membrane separation stage (6) a second permeate stream conduit (12) connected to a permeate gas outlet of the retentate stream membrane separation stage (6) and to the gas-stream conduit (1) at a second gas recycle point (20) or connected to a permeate gas outlet of the retentate stream membrane separation stage (6) and to the fourth retentate stream conduit (18b) at a fourth gas recycle point, located upstream of the main compressor (2) and downstream of the pressure control unit (22), a second retentate stream conduit (13) connected to a retentate gas outlet of the retentate stream membrane separation stage (6), the second retentate stream conduit (13) preferably comprises a pressure control unit (23), a recompressed first permeate stream conduit (14) connected to a gas outlet of the support compressor (9) and to a gas inlet of the permeate stream membrane separation stage (7) a third permeate stream conduit (15) connected to a permeate gas outlet of the permeate stream membrane separation stage (7), the third permeate conduit (15) is preferably configured to discharge the third permeate to the surrounding atmosphere; a third retentate stream conduit (16) connected to a retentate gas outlet of the permeate stream membrane separation stage (7) and to a gas inlet of the fourth membrane separation stage (8) a fourth permeate stream conduit (17) connected to a permeate gas outlet of the fourth membrane separation stage (8) and to the first permeate stream conduit (10) at a third gas recycle point (21) located upstream of the support compressor (9) the first gas recycle point (19) and the second gas recycle point (20), which may be combined to one single gas recycling point, are located upstream of the gas inlet of the main compressor (2) such that the raw gas from the raw gas source, the recycled second permeate stream and the recycled fourth retentate stream are combined to form a gas stream being fed via the gas stream conduit (1) to the gas inlet of the main compressor (2) characterized in that a fourth retentate stream conduit (18a) is connected to a retentate gas outlet of the fourth membrane separation stage (8) and to the gas inlet of the pressure control unit (22), a fourth retentate stream conduit (18b) is connected to the gas outlet of the pressure control unit (22) and to the gas stream conduit (1) at a first gas recycle point (19) the main compressor (2) has a higher compressor load than the support compressor (9) such that the ratio of the power consumption of the support compressor (9) to the power consumption of the main compressor (2) CpOwer(9) / CpOwer(2) is in a range of from 0.001 to 0.8, preferably 0.005 to 0.6, more preferred 0.01 to 0.5, even more preferred 0.01 to 0.4, particularly preferred 0.01 to 0.3 and most preferred 0.01 to 0.25; the support compressor (9) is configured such that the ratio of its outlet pressure to its inlet pressure psupport-out / psupport-in is in a range of from 1 .05 to 3.5, preferably 1.1 to 3, more preferred 1.1 to 2.5 and most preferred 1 .1 to 2.2;the main compressor (2) and the support compressor (9) are configured such that the ratio of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) Pmain-out I Psupport-out>1 the membranes used in separation stages 1 , 2 and 3 are membranes having a higher mixed gas permeability for carbon dioxide than for methane and having a mixed gas selectivity SCO2 / CH4 of 25 to 100, preferably 30 to 100, more preferred 30 to 90, even more preferred 35 to 90, particular preferred 40 to 80 and most preferred 45 to 70, mixed gas permeabilities and mixed gas selectivities are measured at the standard pressure and standard temperature of the respective membrane separation unit; the membranes used in separation stage 4 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 20 to 100, preferably 25 to 90, more preferred 25 to 80 and most preferred 25 to 70, mixed gas permeabilities and mixed gas selectivities are measured at the standard pressure and standard temperature of the respective membrane separation unit.

[0069] It is preferred to connect the outlet of the second retentate stream conduit (13) to a pressurized gas network, more preferred a natural gas grid, even more preferred a CBG (compressed biogas) or LBG (liquefied biogas) network, or to a further processing unit. Preferred further processing units are selected from the group consisting of adsorber unit, absorber unit, temperature swing adsorption, pressure swing adsorption, cryogenic unit, compressor, chiller or tank.

[0070] It is further preferred to arrange a pressure control unit (23) in the second retentate stream (13) and / or a pressure control unit (24) in the second permeate stream (12) and / or a pressure control unit (25) in the fourth permeate stream (17). and / or a pressure control unit (26) in the first retentate stream (11) and / or a pressure control unit (27) in the third retentate stream (16)

[0071] Preferably used pressure control units (22), (23) (24), (25), (26) and (27) can be identical or different and are preferably selected independently from each other from the group of devices consisting of pressure control valves, gate valves, globe valves, ball valves, check valves(spring return valves), butterfly valves, angle valves, ball / plug valves and diaphragm valves.

[0072] In a preferred embodiment the pressure of the raw gas stream (RGS) is in a range of from 0 to 1 barg, preferably 0 to 0.5 barg, more preferred 0.01 to 0.4 barg, even more preferred 0.01 to 0.3 barg and most preferred 0.05 to 0.3 barg.Examples

[0073] The examples which follow serve to provide more particular elucidation and better understanding of the present invention, but do not limit it in any way.Example 1 (inventive)

[0074] Calculations for inventive examples were carried out for gas separation in a facility as shown in Fig. 1 , using a customized model in Aspen Custom Modeler ® process simulation software which calculates mass transfer of the gas components through the membrane by numerical integration of differential equations for mass transfer through a membrane by a solution-diffusion process according to Scholz, M., 2013. Membrane based biogas upgrading processes. Doctoral dissertation, Rheinisch- Westfalische Technische Hochschule Aachen, based on experimental data for the permeance of the membrane for methane and carbon dioxide. All pressures are given as absolute pressure (bara).

[0075] Gas separation was calculated for separating a raw gas stream (RGS) and agas stream (1) having a pressure pgsi= 1 .00 bara. The raw gas stream consisting of CcH4-raw = 55 mol% of methane and Cco2-raw = 45 mol% of carbon dioxide was mixed with recycling streams (12) and (18b) to obtain a gas stream (1) consisting of CcH4-gsi = 50 mol% of methane and Cco2-gsi = 50 mol% of carbon dioxide. The feed stream (3) obtained after compression of the gas stream (1) has the same composition as the gas stream (1) and was separated in a membrane separation unit (4) having a total separation capacity to raw gas flow of 248 GPU m2h / Nm3(referring to the fastest component, i.e., CO2). In the feed stream separation stage (5) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.19 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 , in the retentate separation stage (6) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.33 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 =50.1 , in the permeate separation stage (7) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.22 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 and in the fourth stage (8) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.26 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 25.0 were used.

[0076] Temperature of all four membrane separation stages was set to 25 °C and feed stream (3) pressure was set to pmain-out = 16 bara. Calculations were carried out for isothermal separation neglecting pressure drop on the retentate side of a module. An optimization was carried out with the boundary conditions of providing a methane content of CcH4-ret2 = 98 mol% in the first product stream, i.e. theT1 second retentate stream (13), and a methane yield (methane recovery) with the first product stream, i.e. second retentate stream (13) of 99.93 %. This leads to a methane content of 0.18 mol% in the second product stream, i.e. the third permeate stream (15).

[0077] Stage cuts in all four stages, pressure ratio over the support compressor (9) Psupport-out / Psupport-in, permeate side pressure of the first and forth membrane separation (pPermi and pPerm4) were varied to minimize power consumption and membrane count, ensuring a good compromise between both operational and investment costs. The optimization calculated for the optimum configuration a pressure ratio over the support compressor (9) psupport-out / psupport-in = 4.22 bara / 2.02 bara = 2.09. The calculated power consumption of the main compressor (2) was CpOwer(2) = 54.8 kW and the power consumption of the support compressor (9) was CpOwer(9) = 7.6 KW. Thus, CpOwer(9) / CpOwer (2) = 0.14. The optimization calculated the following stage cuts SC1 = 0.39, SC2 = 0.28, SC3 = 0.60 and SC4 = 0.80.

[0078] The calculated flow rates, pressures and compositions of the different gas streams are given inTable 9.

[0079] Table 9

[0080] Example 1 shows that the process and facility of the invention can upgrade a typical CO2 / CH4 gas mixture to biomethane having a methane content of 98 mol% with a very high methane recovery of 99.93 % with very low energy demand, i.e. operational costs. The process of the invention separates a major part of the carbon dioxide with a gas stream (15) containing only 0.18 mol% of methane which can be discharged directly to the atmosphere even at locations with highest environmental standards. No post treatment of the off-gas streams is necessary.Example 2 (inventive)

[0081] Example 1 was repeated with the following modifications:

[0082] A facility (4) having a total separation capacity to raw gas flow of 238 GPU m2h / Nm3was used. In the feed stream separation stage (5) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.20 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 , in the retentate separation stage (6) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.34 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 , in the permeate separation stage (7) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.23 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 and in the fourth stage (8) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.24 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 were used. Thus, compared to Example 1 , a higher selective membrane was used in the fourth stage (8).

[0083] Volume flow rates, compositions and pressures of the two product streams, i.e. the second retentate stream (13) and the third permeate stream (15) were set identical to Example 1 . The raw gas stream and the temperature of all four stages were set identical to Example 1 , too. Thus, comparison of Example 1 and 2 shows the technical effect of modifying the selectivity SCO2 / CH4 of the membranes used in stage 4.

[0084] Stage cuts in all four stages, pressure ratio over the support compressor (9) Psupport-out / Psupport-in, permeate side pressure of the first and forth membrane separation (pPermi and pperm4) were varied to minimize power consumption and membrane count, ensuring a good compromise between both operational and investment costs. The optimization calculated a pressure ratio over the support compressor (9) psupport-out / psupport-in = 4.32 bara / 1 .99 bara = 2.17. The calculated power consumption of the main compressor (2) was CpOwer (2> = 54.4 kW and the power consumption of the support compressor (9) Was Cpower (9) — 7.7 KW. Thus, Cpower(9) / Cpower (2)—0.14. Thus, use of a higher selective membrane in the fourth separation stage (8) slightly decreases the operational costs because of the slightly lower power consumption, whereas depending on commercial membrane material pricing it might lead to an increase in the invest. Based on the location of the plant and the focus on operational or invest costs, the process and apparatus of the invention can be adjusted accordingly. The optimization calculated the following stage cuts SC1 = 0.39, SC2 = 0.28, SC3 = 0.63 and SC4 = 0.79.

[0085] The calculated flow rates, pressures and compositions of the different gas streams are given in Table 10.

[0086] Table 10

[0087] Example 2 shows that the process of the invention even with the modified configuration can upgrade a typical CO2 / CH4 gas mixture to biomethane having a methane content of 98 mol% with a methane recovery of 99.93 % with very low energy demand. The process of the invention separates a major part of the carbon dioxide with a gas stream (15) containing only 0.18 mol% of methane which can be discharged directly to the atmosphere even at locations with highest environmental standards. No post treatment of the off-gas streams is necessary. Use of membranes with higher selectivity SCO2 / CH4 in stage 4 has shown that it can allow to flexibly optimize costs.Comparison Example 1 (non-inventive)

[0088] Calculations were carried out for gas separation in a facility as shown in Fig. 2, i.e. the facility disclosed in EP2735355A1 , Figure 5, using process simulation software which calculates mass transfer of the gas components through the membrane by numerical integration of the known differential equations for mass transfer through a membrane by a solution-diffusion process, based on experimental data for the permeance of the membrane for methane and carbon dioxide. All pressures are given as absolute pressure.

[0089] The configuration according to EP2735355A1 , Figure 5, differs from the inventive process and facility in that the fourth retentate stream is combined with the second retentate stream and withdrawn as mixed first product stream. In contrast thereto, in the present invention the fourth retentate stream (18) is recycled back to the up-stream side of the main compressor (2). The configuration according toEP2735355A1 , Figure 5 requires that support compressor C2 provides a compressed first permeate stream having a pressure that is similar to or higher than the second retentate stream (1 1). Otherwise the fourth retentate stream (18) cannot be fed to the second retentate stream (11).

[0090] A facility in Comparison Example 1 consisting of a membrane separation unit (4) having a total separation capacity to raw gas flow of 153 GPU m2h / Nm3referring to the fastest component and scaled over raw gas flow was used. In the feed stream separation stage (5) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.32 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 , in the retentate separation stage (6) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.08 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 , in the permeate separation stage (7) membranes with a ratio of total capacity over stage to total capacity of the membrane separation unit (4) of 0.08 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 and in the fourth stage (8) membranes with a ratio of total capacity of per stage to total capacity of the membrane separation unit (4) of 0.52 having a selectivity for carbon dioxide over methane under standard conditions of SCO2 / CH4 = 50.1 were used. Thus, the membrane selectivities SCO2 / CH4 in the four stages are the same as in Example 2.

[0091] Volume flow rates, compositions and pressures of the two product streams, i.e. the second retentate stream and the third permeate stream in Comparison Example 1 were set identical to Examples 1 and 2. The raw gas stream used in Comparison Example 1 was identical to Examples 1 and 2, too. Also the temperature in all four stages. Thus, comparison of Comparison Example 1 to Examples 1 and 2 shows the technical effect of the modified use of the fourth retentate stream.

[0092] Stage cuts in all four stages, pressure ratio over the support compressor (9) Psupport-out / Psupport-in, permeate side pressure of the first and forth membrane separation (pPermi and pperm4) were varied to minimize power consumption and membrane count, ensuring a good compromise between both operation costs and investment costs. The optimization calculated a pressure ratio over the support compressor (9) psupport-out / psupport-in = 16.00 bara / 2.3 bara = 6.96. The power consumption of the main compressor (2) CpOwer(2) = 52.0 kW and the power consumption of the support compressor (9) CpOwer(9) = 19.0 KW. Thus, Cpower(9) / Cpower (2) = 0.37. The optimization calculated the following stage cuts SC1 = 0.39, SC2 = 0.29, SC3 = 0.66 and SC4 = 0.87.

[0093] The calculated flow rates, pressures and compositions of the different gas streams are given inTable 11.

[0094] Table 11

[0095] Even though the double compression rate of the main compressor (2) is lower in Comparative Example 1 (21 .4%) compared to Example 1 (28.4%) and Example 2 (27.3%), the overall energy consumption of the inventive facility is lower. This is because the inventive use of the fourth retentate stream allows to operate separation stages 3 and 4 at significantly lower retentate pressures than required in Comparative Example 1. As consequence, the required recompression rate of the support compressor (9) in Comparative Example 1 (2.3 bara to 16 bara) is significantly higher than in Example 1 (2.02 bara to 4.22 bara) and Example 2 (1 .99 bara to 4.32 bara). Thus, the savings in energy demand of the support compressor (9) in the inventive facility overcompensates the slightly higher energy demand of the main compressor (2) compared to Comparative Example 1 . In addition, a smaller compressor can be used in the inventive process and facility, which leads to additional savings.

[0096] This shows that the process and facility of the invention has significant benefits compared to the prior art configuration of EP2735355A1 .Example 3

[0097] Example 3 shows the effects of variations of the raw gas composition and the membrane selectivity SCO2 / CH4 on the ratio of total power consumption to raw gas flow, which is used as representative parameter for the operational costs.

[0098] The ratio of total power consumption to raw gas flow for a facility according to the invention was compared with the ratio of total power consumption to raw gas flow for a facility according to the prior art US 7537641 B2, Figure 2.

[0099] The simulations underlying this comparison were conducted under the premises that the following boundary conditions were varied three different CO2 concentrations in the raw gas stream (RGS) of Cco2-raw of 30 mol%, 40 mol% and 50 mol% were simulated, and three different mixed gas selectivities SCO2 / CH4 at standard conditions of the respective membrane separation unit of SCO2 / CH4 = 30, SCO2 / CH4 = 50 and SCO2 / CH4 = 70 were simulated. In all simulations identical membranes with identical selectivities SCO2 / CH4 were used in all four membrane separation units (5), (6), (7) and (8) and the following boundary conditions were kept constant the raw gas stream further comprises 0.5 mol% O2, the rest is CH4, the methane concentration CcH4-ret2 in the second retentate stream (13) is set to 97 mol% the pressure pmain-out at the gas outlet of the main compressor (2) is set to 10 bara The O2 / CH4 selectivity was fixed to S02 / CH4 10, methane recovery was set to 99.5 %

[0100] A large number of simulations were carried out where the pressure ratio over the support compressor (9) psupport-out / psupport-in , the pressure of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17) which was set equal to (10), the stage cut SC2 of the retentate stream membrane separation stage (6) and the stage cut SC4 of the fourth membrane separation stage (8) were varied. The pressure ratio over the support compressor (9) psupport-out / psupport-in was varied between 1.1 and 2.2, the pressure of the first permeate stream (10) (and fourth permeate stream (17), respectively) was varied between 1 bara and 1 .2 times the square root of the feed pressure to stage 1 , the stage cut SC2 of the retentate stream membrane separation stage (6) was varied within 0.08 and 0.58 mol / mol and the stage cut of the SC4 fourth membrane separation stage (8) between 0.1 and 0.9 mol / mol. The ranges for the aforementioned pressure ratio and permeate pressure were chosen based on previous experience and tests. The ranges for the aforementioned stage cuts were chosen to promote convergence of the simulation given the specified boundary conditions. Only the simulations leading to the minimum values for the total power consumption (as a sum of that of the main (2) and the support compressor (9)), for each possible combination of specified values for the boundary conditions, were taken for the comparison. The minimum values for the total power consumption to raw gas flow for each case are shown in Figure 3. The three filled square symbols show the simulation results for SCO2 / CH4 = 70 and either Cco2-raw = 30 mol%, 40 mol% or 50 mol % respectively. The three "x" symbols show the simulation results for SCO2 / CH4 = 50 and either Cco2-raw = 30 mol%, 40 mol% or 50 mol % respectively. The three filled circles show the simulation results for SCO2 / CH4 = 30 and either Cco2-raw = 30 mol%, 40 mol% or 50 mol % respectively. An arrow in Figure 3 indicates increasing CO2 content in the raw gas stream. Each of these nine symbols show the ratio of total power consumption to raw gas flow where all premises defined above are met with the inventive facility and process as well as with the facility and process according to the prior art US 7537641 B2, Figure 2.

[0101] As can be seen in Figure 3, in all simulated cases the ratio of total power consumption to raw gas flow of the facility and process of the present invention is lower compared to US 7537641 B2. The ratio oftotal power consumption to raw gas flow correlates to the operational costs of the processes and devices which are significantly lower for the facility and process of the invention.

[0102] The energy saving achieved with the facility and process of the invention compared to US 7537641 B2 is higher for higher CO2 concentrations in the raw gas stream (see arrow in Figure 3). Figure 3 further shows that with the process and device of the invention the ratio of total power consumption to raw gas flow is nearly independent from the CO2 concentrations in the feed stream. Thus, facility and process of the invention can be used for varying raw gas compositions. In contrast thereto, the ratio of total power consumption to raw gas flow of the process and device of US 7537641 B2, Figure 2 are strongly dependent from the CO2 concentrations in the raw gas stream. In the process and device of US 7537641 B2, Figure 2, operational costs increase if the CO2 concentration in the raw gas stream is increased.

[0103] Figure 3 further shows that the ratio of total power consumption to raw gas flow and thus the operational costs of the facility and process of the invention can be reduced if higher selective membranes are used.

[0104] Figure 3 finally confirms that the process and device of the invention can be used to treat raw gas streams with different compositions but nevertheless achieves very high methane recovery, high methane purity and low operating costs.Example 4

[0105] Example 4 shows the effects of variations of the pressure pmain-out generated by the compressor (2) and the membrane selectivity SCO2 / CH4 on the ratio of total power consumption to raw gas flow, which is used as representative parameter for the operational costs. Flexibility concerning Pmain-out is required to allow different use opportunities of the first product stream, i.e. the second retentate stream (13), without use of an additional compressor in said stream.

[0106] As in Example 3, the ratio of total power consumption to raw gas flow for a facility according to the invention was compared with the ratio of total power consumption to raw gas flow for a facility according to the prior art US 7537641 B2, Figure 2.

[0107] The simulations underlying this comparison were conducted under the premises that that the following boundary conditions were varied five different pressures pmain-out at the gas outlet of the main compressor (2) of 8 bara, 10 bara, 12 bara, 14 bara and 16 bara were simulated, three different mixed gas selectivities SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of SCO2 / CH4 = 30, SCO2 / CH4 = 50 and SCO2 / CH4 = 70 were simulated. In all simulations membrane selectivities SCO2 / CH4 are set identical in all four membrane separation units (5), (6), (7) and (8) and the following boundary conditions were kept constantthe CO2 concentration in the raw gas stream (RGS) of Cco2-raw is set to 40 mol% and the raw gas further comprises 0.5 mol % O2, the rest is CH4, the methane concentration CcH4-ret2 in the second retentate stream (13) is set to 97 mol% the O2 / CH4 selectivity was fixed to S02 / CH4 =10, methane recovery was set to 99.5 %

[0108] As in Example 3, a large number of simulations were carried out where the pressure ratio over the support compressor (9) psupport-out / psupport-in, the pressure of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17) which was set equal to that of stream (10), the stage cut SC2 of the retentate stream membrane separation stage (6) and the stage cut SC4 of the fourth membrane separation stage (8) were varied. The pressure ratio over the support compressor (9) Psupport-out I psupport-in was varied between 1.1 and 2.2, the pressure of the first permeate stream (10) (and fourth permeate stream (17), respectively) was varied between 1 bara and 1 .2 times the square root of the feed pressure to stage 1 , the stage cut of stage 2 was varied within 0.08 and 0.58 mol / mol and the stage cut of stage 4 between 0.1 and 0.9 mol / mol. The ranges for the aforementioned pressure ratio and permeate pressure were chosen based on previous experience and tests. The ranges for the aforementioned stage cuts were chosen to promote convergence of the simulation given the specified boundary conditions.Only the simulations leading to the minimum values for the total power consumption (as a sum of that of the main and the support compressor), for each possible combination of specified values for the boundary conditions, were taken for the comparison. The minimum values for the total power consumption to raw gas flow for each case are shown in Figure 4. The five filled square symbols show the simulation results for a first series with SCO2 / CH4 = 70 and either pmain-out = 8 bara, 10 bara, 12 bara, 14 bara or 16 bara respectively. The five “x” symbols show the simulation results for a second series with SCO2 / CH4 = 50 and either pmain-out = 8 bara, 10 bara, 12 bara, 14 bara or 16 bara respectively. The five filled circles show the simulation results for a third series with for SCO2 / CH4 = 30 and either pmain-out = 8 bara, 10 bara, 12 bara, 14 bara or 16 bara respectively. An arrow In Figure 4 indicates increasing pmain-out for each series. Each of these 15 symbols show the ratio of total power consumption to raw gas flow where all premises defined for Example 4 above are met with the inventive facility and process as well as with the facility and process according to the prior art US 7537641 B2, Figure 2.

[0109] As can be seen in Figure 4, in all simulated cases the ratio of total power consumption to raw gas flow of the facility and process of the present invention is lower compared to US 7537641 B2 if the same raw gas is treated and the product gas specifications are the same. Thus, the device and process of the invention can be effectively used to provide methane enriched product gas streams (13) that meet different pressure requirements, for example of gas pipelines, at lower operational costs but with constantly high methane purity and recovery.

[0110] Figure 4 further shows that the operational costs (represented by the ratio of total power consumption to raw gas flow of the facility) and investment costs (represented by the fact that higher selective membranes have lower permeance and therefore typically lead to higher costs for a given separation task) can be optimized. For example, a pmain-out of 8 bara can be achieved at a ratio of total power consumption to raw gas flow of the facility below 0.1 with membranes having SCO2 / CH4 = 50 while with the process and facility according to US 7537641 B2, Figure 2, even with membranes requiring ahigher total separation capacity having SCO2 / CH4 = 70, a pmain-out of 8 bara can only be achieved at a ratio of total power consumption to raw gas flow of the facility of more than 0.1 . Thus, the optimum for a preset Pmain-out of 8 bara for the process and facility of the present invention can be reached at lower operational and investment costs compared to US 7537641 B2, Figure 2.Example 5 (inventive)

[0111] Example 5 shows the effects of variations of the required product purity CcH4-ret2 and the membrane selectivity SCO2 / CH4 on the ratio of total power consumption to raw gas flow, which is used as representative parameter for the operational costs. Flexibility concerning CcH4-ret2 is required to allow different use opportunities of the first product stream, i.e. the second retentate stream (13).

[0112] As in Example 3, the ratio of total power consumption to raw gas flow for a facility according to the invention was compared with the ratio of total power consumption to raw gas flow for a facility according to the prior art US 7537641 B2, Figure 2.

[0113] The simulations underlying this comparison were conducted under the premises that that the following boundary conditions were varied six different methane concentration CcH4-ret2 in the second retentate stream (13) of 97 mol%, 97.5 mol%, 98 mol%, 98.5 mol%, 99 mol% and 99.5 mol% were simulated three different mixed gas selectivities SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of SCO2 / CH4 = 30, SCO2 / CH4 = 50 and SCO2 / CH4 = 70 were simulated. In all simulations membrane selectivities SCO2 / CH4 are set identical in all four membrane separation units (5), (6), (7) and (8) and the following boundary conditions were kept constantCO2 concentration in the raw gas stream (RGS) Cco2-raw was set to 40 mol%, the raw gas further comprises 0.5 mol % O2, the rest is CH4, the pressure pmain-out at the gas outlet of the main compressor (2) is set to 10 bara the O2 / CH4 selectivity was fixed to 10, methane recovery was set to 99.5 %

[0114] As in Examples 3 and 4, a large number of simulations were carried out where the pressure ratio over the support compressor (9) psupport-out / psupport-in, the pressure of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17) which was set equal to that of stream (10), the stage cut SC2 of the retentate stream membrane separation stage (6) and the stage cut SC4 of the fourth membrane separation stage (8) were varied. The pressure ratio over the support compressor (9) Psupport-out I psupport-in was varied between 1.1 and 2.2, the pressure of the first permeate stream (10) (and fourth permeate stream (17), respectively) was varied between 1 bara and 1 .2 times the square root of the feed pressure to stage 1 , the stage cut SC2 of the retentate stream membrane separation stage (6) was varied within 0.08 and 0.58 mol / mol and the stage cut SC4 of the fourth membrane separation stage (8) between 0.1 and 0.9 mol / mol. The ranges for the aforementioned pressure ratio and permeate pressure were chosen based on previous experience and tests. The ranges for the aforementioned stage cuts were chosen to promote convergence of the simulation given the specified boundary conditions. Only the simulations leading to the minimum values for the total power consumption (as a sum of that of the mainand the support compressor), for each possible combination of specified values for the boundary conditions, were taken for the comparison. The minimum values for the total power consumption to raw gas flow for each case are shown in Figure 5. The six filled square symbols show the simulation results for a first series with SCO2 / CH4 = 70 and either CcH4-ret2 = 97 mol%, 97.5 mol%, 98 mol%, 98.5 mol%, 99 mol% or 99.5 mol% respectively. The six “x” symbols show the simulation results for a second series with SCO2 / CH4 = 50 and either CcH4-ret2 = 97 mol%, 97.5 mol%, 98 mol%, 98.5 mol%, 99 mol% or 99.5 mol% respectively. The six filled circles show the simulation results for a third series with SCO2 / CH4 = 30 and either CcH4-ret2 = 97 mol%, 97.5 mol%, 98 mol%, 98.5 mol%, 99 mol% or 99.5 mol% respectively. An arrow in Figure 5 indicates increasing CcH4-ret2 for each series. Each of these 18 symbols show the ratio of total power consumption to raw gas flow where all premises defined for Example 5 above are met with the inventive facility and process as well as with the facility and process according to the prior art US 7537641 B2, Figure 2.

[0115] As can be seen in in Figure 5, in all simulated cases the ratio of total power consumption to raw gas flow of the facility and process of the present invention is lower compared to US 7537641 B2 if the same raw gas is treated and the product gas specifications are the same. Thus, the device and process of the invention can be effectively used to provide methane enriched product gas streams (13) that meet different requirements regarding methane purity, at lower operating costs but with constantly high methane recovery. Even very pure methane streams can be obtained.

[0116] Figure 5 further shows that the operating costs (represented by the ratio of total power consumption to raw gas flow of the facility) and investment costs (represented by the fact that higher selective membranes have lower permeance and therefore typically lead to higher costs for a given separation task) can be optimized. For example, a CcH4-ret2 = 97 mol% can be achieved at a ratio of total power consumption to raw gas flow of the facility below 0.105 with membranes having SCO2 / CH4 = 50 while with the process and facility according to US 7537641 B2, Figure 2, even with a more expensive membrane having SCO2 / CH4 = 70, a CcH4-ret2 = 97 mol% can only be achieved at a ratio of total power consumption to raw gas flow of the facility of more than 0.1 1 . Thus, the optimum for a pre-set CcH4-ret2 = 97 mol% for the process and facility of the present invention can be reached at lower operational and investment costs compared to US 7537641 B2, Figure 2.

[0117] Figure 5 further shows that the total energy consumption of the process and device of the invention can be optimized by selection of the membrane selectivity SCO2 / CH4. Higher selective membranes lead to a reduced total energy consumption if the same raw gas is treated, and the product gas specifications are the same.Examples 6 to 8

[0118] Examples 3 to 5 were repeated in each case with an increased methane recovery of 99.9% instead of 99.5%. All possible combinations of boundary conditions for the mixed gas selectivities SCO2 / CH4, CO2 concentration Cco2-raw in the raw gas stream (RGS), pressure pmain-out at the gas outlet of the main compressor (2) and methane concentration CcH4-ret2 in the second retentate stream (13), in the same intervals as in examples 3 to 5 were studied. As in Examples 3 to 5, a large number of simulations werecarried out where the pressure ratio over the support compressor (9) psupport-out / psupport-in , the pressure of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17) which was set equal to), the stage cut SC2 of the retentate stream membrane separation stage (6) and the stage cut SC4 of the fourth membrane separation stage (8) were varied. The pressure ratio over the support compressor (9) psupport-out / psupport-in was varied between 1.1 and 2.2, the pressure of the first permeate stream (10) (and fourth permeate stream (17), respectively) was varied between 1 bara and the square root of the feed pressure to stage 1 , the stage cut SC2 of the retentate stream membrane separation stage (6) was varied within 0.1 and 0.9 mol / mol and the stage cut SC4 of the fourth membrane separation stage (8) between 0.4 and 0.9 mol / mol. The ranges for the aforementioned pressure ratio and permeate pressure were chosen based on previous experience and tests. The ranges for the aforementioned stage cuts were chosen to promote convergence of the simulation given the specified boundary conditions. Only the simulations leading to the minimum values for the total power consumption (as a sum of that of the main and the support compressor), for each possible combination of specified values for the boundary conditions, were taken for the comparison.

[0119] The minimum values for the total power consumption to raw gas flow for each case are summarized in Figure 6. Each filled square, "x" and filled circle shows the total power consumption of the process and apparatus of the invention as well as of the process and apparatus of US 7537641 B2 for one specific process configuration, i.e. one possible combination of boundary conditions (e.g. Cco2-raw = 40%, Product purity, CcH4-ret2 = 98% mol, methane recovery = 99.9% , Pmain-out — 16 bara, SCO2 / CH4 = 50, represents one process configuration) chosen from the set of values defined in Examples 3 to 5. As can be seen in Figure 6, in all simulated cases the ratio of total power consumption to raw gas flow of the facility and process of the present invention is lower than that of the device according to US 7537641 B2 if the same raw gas is treated and the specifications of the first product gas are the same. Thus, the device and process of the invention can be effectively operated with reduced operational and investment even at locations with very strict requirements regarding methane emissions. Even highest environmental standards can be met.Example 9 (inventive)

[0120] Examples 1 to 8 demonstrate that the process and facility of the invention have reduced operational costs compared to the prior art processes and facilities according to US 7537641 B2 and EP2735355A1 . In Example 9 a further assessment of investment costs against operational costs for the process and facility of the invention were done. The joined investment and operational costs assessment was done by comparing the total separation capacity used in the inventive facility (4) (which has an influence on investment costs and the ratio of total power consumption to raw gas flow (which has an influence on the operational costs).

[0121] The simulations underlying this comparison were conducted under the premises that that the following boundary conditions were varied three different CO2 concentrations in the raw gas stream (RGS) of Cco2-raw of 30 mol%, 40 mol% and 50 mol % were simulated,the raw gas further comprises 0.5 mol % O2, the rest is CH4, six different methane concentration CcH4-ret2 in the second retentate stream (13) of 97 mol%, 97.5 mol%, 98 mol%, 98.5 mol%, 99 mol% and 99.5 mol% were simulated five different pressures pmain-out at the gas outlet of the main compressor (2) of 8 bara, 10 bara, 12 bara, 14 bara and 16 bara were simulated three different mixed gas selectivities SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of SCO2 / CH4 = 30, SCO2 / CH4 = 50 and SCO2 / CH4 = 70 were simulated. In all simulations membrane selectivities SCO2 / CH4 are set identical in all four membrane separation units (5), (6), (7) and (8), that the following boundary conditions were kept constant the O2 / CH4 selectivity was fixed to S02 / CH4 =10, methane recovery was set to 99.5 % in a first series of simulations and to 99.9% in a second series of simulations.

[0122] All possible combinations of boundary conditions for the set of values defined, e.g. Cco2-raw = 50%, Product purity, CcH4-ret2 = 99% mol, methane recovery = 99.5% , Pmain-out — 8 bara, SCO2 / CH4 = 50, represents one process configuration, were considered. For all process configurations a large number of simulations were carried out where the pressure ratio over the support compressor (9) Psupport-out / Psupport-in , the pressure of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17) which was set equal to) and the stage cuts of stages 2 and 4 were varied within the same bounds as examples 3 to 5 for yield 99.5%, and within the same bounds as examples 6 to 8 for yield 99.9%. The results for the configuration Cco2-raw = 50%, Product purity, CcH4-ret2 = 99% mol, methane recovery = 99.5%, pmain-out = 8 bara, SCO2 / CH4 = 50 are shown in Figure 7. In Figure 7 open black circles represent data using the facility and process of the present invention.Example 10 (inventive)

[0123] Based on the results of Example 9 further process parameters were regressed to further narrow down inventive process configurations to those having good operational costs (total energy consumption).

[0124] The boundary conditions CcH4-ret2 , Cco2-raw, pmain-out, S02 / CH4 and SCO2 / CH4 were set as described in Example 9. After extensive investigations it was found that it is possible to narrow down the process and facility configurations to those having low ratio of total power consumption to raw flow (influencing the operational costs), if the following additional process parameters were bounded within the following limits: the ratio pmain-out / psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Vam bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure ppermi of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17)) is in a range of from 1 to Var2h bara, wherein Vamti being the upper limit of the pressure range, andthe stage cut SCi of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from Var4i to Var4h mol / mol, wherein Var4i being the lower bound and Var4h being the upper bound of the stage cut range, and the stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from Var5i to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.1 to 0.9 mol / mol in the case of a set methane yield of 99.5% for all values of selectivity studied, and from 0.4 to 0.9 mol / mol in the case of a set methane yield of 99.9% and selectivity SCO2 / CH4 = 50 or above 50, preferably 50 to 70. In the further case of a set methane yield of 99.9% and selectivity SCO2 / CH4 = 30, the range of stage cut SC4 is of from 0.7 to 0.9 mol / mol. The lower bound scales linearly with selectivity between the values 0.7 and 0.4 for selectivity between 30 and 50 (e.g. for selectivity equal to 30 the lower bound is equal to 0.7, for selectivity equal to 40 the lower range is equal to 0.55 and for selectivity equal to 50 the lower range is equal to 0.4) while the upper bound stays at 0.9,

[0125] Intense studies have been conducted to find out that these additional parameters can be used to further optimize the process and facility.

[0126] The bounds Vam, Vanh, Var2h, Varsi, Varsh, Vam, Var4h, Varsi and Vanh, are regressed variables calculated by use of equation (I) above with “Intercept” and “ai” being fixed factors taken from the Table 1 above for a set methane yield of 99.5% and Table 2 for a set methane yield of 99.9% above, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 are values taken from the ranges defined for Example 9 above. The results for the configuration Cco2-raw = 50%, Product purity, CcH4-ret2 = 99% mol, methane recovery = 99.5%, pmain-out = 8 bara, SCO2 / CH4 = 50 after limiting the specified process parameters to the given bounds are shown in Figure 8. In Figure 8 open black circles represent data using the facility and process of the present invention. The configuration in Figure 8 is the same as in Figure 7. Thus, comparison of Figure 8 to Figure 7 shows the effect of the regression to further narrow down inventive process configurations.Example 11 (inventive)

[0127] Based on the results of Example 10 further process parameters were regressed to further narrow down inventive process configurations found in Examples 9 and 10 to those most efficient having the lowest operational costs (total energy consumption).

[0128] The boundary conditions CcH4-ret2 , Cco2-raw, pmain-out, S02 / CH4 and SCO2 / CH4 were set as described in Example 9.

[0129] After intense studies, the inventors found out that the process and facility configurations can be narrowed down to those having the lowest total separation capacity (influencing the investment costs) and the lowest ratio of total power consumption to raw gas flow (influencing the operational costs), if the following additional process parameters were further bounded to the following limits: the ratio pmain-out / psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Vam bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure ppermi of the first permeate stream (10) (and correspondingly that of the fourth permeate stream (17)) is in a range of from 1 bara to Var2h bara, wherein Vamti being the upper limit of the pressure range, and the stage cut SCi of the feed stream membrane separation stage (5) is in a range of from Vam to Vamh mol / mol, wherein Vami being the lower bound and Vamh being the upper bound of the stage cut range, and the CO2 concentration in the fourth retentate stream (18) (Cco2(i8>) has an upper bound of Vareh in mole fraction, and the CO2 concentration in the fourth permeate stream (17) (Cco2(i7>) has an upper bound of Va in mole fraction, and the minimum methane recovery was set to 99.5 %

[0130] The lower and upper bound Vam and Vanh, as well as the upper bound Var2h are regressed variables being calculated by use of equation (II) above, with “ai” being the coefficients for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 3 above, not all possible product pairs or squared terms in equation (II) are significant, therefore only a few coefficients appear in Table 3 while the rest are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 are values taken from the ranges defined in Example 9 above.

[0131] The lower bound Vami and the upper bound Vamh are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 4 above below, not all possible product pairs or squared terms in equation (III) are significant, therefore only a few coefficients appear in Table 4 while the rest is zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw, and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being theoutlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, xs and X4 are values taken from the ranges defined in Example 9 above, and xs is taken from the ranges defined by Vam and Vanh in the current example (with the corresponding conversion using the value of X2).

[0132] Vareh and Var?h are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “ai»2” being fixed factors taken from the Table 5 above, not all possible product pairs or squared terms in equation (III) are significant, therefore only a few coefficients appear in Table 5 while the rest is zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw, Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 — Psupport-out, and wherein xi , X2, xs and X4 are values taken from the ranges defined in Example 9 above, and xs is taken from the ranges defined by Vam and Vanh in the current example (with the corresponding conversion using the value of x2). The results for the configuration Cco2-raw = 50%, Product purity, CcH4-ret2 = 99% mol, methane recovery = 99.5% , Pmain-out — 8 bara, SCO2 / CH4 = 50 after further limiting the specified process variables to the given bounds are shown in Figure 9. In Figure 9 open black circles represent data using the facility and process of the present invention.Example 12 (inventive)

[0133] Example 11 was repeated but with a set minimum methane recovery set to 99.9 % instead of 99.5%.

[0134] The lower and upper bound Vam and Vanh, as well as the upper bound Var2h are regressed variables being calculated by use of equation (II) with “ai” being the coefficients for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 6 above, not all possible product pairs or squared terms in equation (II) are significant, therefore only a few coefficients appear in Table 6 while the rest are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined in Example 9 above.

[0135] The bounds Vam and Varsh are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “ai ” and “a^” being fixed factors taken from the Table 7 above, not all possible product pairs or squared terms in equation (III) are significant, therefore only a few coefficients appear in Table 7 while the rest are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2 , Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = pmain-out, Xi with i= 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw, Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi, X2, xs and X4 are values taken from the ranges defined in Example 9, and xs is taken from the ranges defined by Vam and Vanh in the current example (with the corresponding conversion using the value of X2).

[0136] Vareh and Var?h are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “ai»2” being fixed factors taken from the Table 8 above, not all possible product pairs or squared terms in equation (III) are significant, therefore only a few coefficients appear in Table 8, while the rest are zero), Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw, Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = Psupport-out, and wherein xi , X2, X3 and X4 are values taken from the ranges defined in Example 9, and X5 is taken from the ranges defined by Vam and Vanh in the current example (with the corresponding conversion using the value of X2).Example 13 (inventive)

[0137] Calculations for Example 2 were repeated with a typical biogas composition. The composition of the raw gas consisted of CcH4-raw = 55.0 mol% of methane, CN2-raw = 0.1 mol% of nitrogen, Co2-raw = 0.4 mol% of oxygen and Cco2-raw = 44.5 mol% of carbon dioxide.

[0138] A facility (4) having a total separation capacity to raw gas flow of 316 GPU m2h / Nm3was used. In the feed stream separation stage (5) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.18, in the retentate separation stage (6) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.29, in the permeate separation stage (7) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.18 and, in the fourth stage (8) membranes with a ratio of total capacity per stage to total capacity of the membrane separation unit (4) of 0.35 were used. All four stages were having a fixed selectivity for carbon dioxide over methane of SCO2 / CH4 = 50.1 , a fixed selectivity for nitrogen over methane of SN2 / CH4 = 1.6 and a fixed selectivity for oxygen over methane of S02 / CH4 = 10.0, all under standard conditions.

[0139] Volume flow rates, compositions of methane and pressures of the two product streams, i.e. , the second retentate stream (13) and the third permeate stream (15) were set identical to Example 2. The volume flow rate and pressure of the raw gas, as well as the temperature of all four stages, were also set identical to Example 2.

[0140] Stage cuts in all four stages, pressure ratio over the support compressor (9) Psupport-out I Psupport-in, permeate side pressure of the first and forth membrane separation (pPermi and pperm4) were varied tominimize power consumption and membrane count, ensuring a good compromise between both. The optimization calculated a pressure ratio over the support compressor (9) psupport-out / psupport-in = 4.19 bara / 2.33 bara = 1 .8. The calculated power consumption of the main compressor (2) was Cpower (2) = 55.5 kW and the power consumption of the support compressor (9) was Cp0Wer (9) = 6.9 KW. Thus, COwer(9) / COwer (2) = 0.12. Compared to Example 2, therefore, the present example only leads to a slight increase in operational costs, shown by the slightly higher power consumption, and to some increase in investment costs, due to the extra requirements on the total separation capacity.

[0141] The calculated flow rates, pressures and compositions of the different gas streams are given in Table 12.Table 12

[0142] Reference numbers(1) gas-stream respectively gas-stream conduit (1)(2) main compressor(3) feed stream respectively feed stream conduit(4) membrane separation unit(5) feed stream membrane separation stage(6) retentate stream membrane separation stage(7) permeate stream membrane separation stage(8) fourth membrane separation stage(9) support compressor(10) first permeate stream respectively first permeate stream conduit(11) first retentate stream respectively first retentate stream conduit(12) second permeate stream respectively second permeate stream conduit(13) second retentate stream respectively second retentate stream conduit(14) recompressed first permeate stream respectively recompressed first permeate stream conduit(15) third permeate stream respectively third permeate stream conduit(16) third retentate stream respectively third retentate stream conduit(17) fourth permeate stream respectively fourth permeate stream conduit(18) fourth retentate stream respectively fourth retentate stream conduit(18a / b) fourth retentate stream respectively fourth retentate stream conduit up-stream (18a) respectively downstream (18b) of the pressure reduction unit vale (22)(19) first gas recycle point(20) second gas recycle point(21) third gas recycle point(22) pressure control unit in the fourth retentate stream (18); in Figure 1 shown as pressure control valve(23) pressure control unit second retentate stream (13); in Figure 1 shown as pressure control valve(24) pressure control unit in the second permeate stream (12) (for simplification of the drawing not shown in Figure 1)(25) pressure control unit (25) in the fourth permeate stream (17) (for simplification of the drawing not shown in Figure 1)(26) pressure control unit (26) in the first retentate stream (11) (for simplification of the drawing not shown in Figure 1)(27) pressure control unit (27) in the third retentate stream (16) (for simplification of the drawing not shown in Figure 1)

Claims

Claims1 . A process for separating methane from a gas stream comprising methane and carbon dioxide, comprising the following steps:(a) providing a gas-stream (1), comprising the raw gas stream (RGS) comprising methane and carbon dioxide(b) compressing the gas stream (1) with a main compressor (2) to provide a feed stream (3)(c) passing the feed stream (3) to a membrane separation unit (4) comprising a feed stream membrane separation stage (5), a retentate stream membrane separation stage (6), a permeate stream membrane separation stage (7), a fourth membrane separation stage (8), and a support compressor (9);(d) passing the feed stream (3) to a feed gas inlet of the feed stream membrane separation stage (5) and processing the feed stream (3) in the feed stream membrane separation stage (5) to produce a first permeate stream (10) at lower pressure than the feed stream, which is enriched in carbon dioxide compared to the feed stream (3), and a first retentate stream (11), which is enriched in methane compared to the feed stream (3);(e) passing the first retentate stream (1 1) to the gas inlet of the retentate stream membrane separation stage (6) and processing the first retentate stream (11) in the retentate stream membrane separation stage (6) to produce a second permeate stream (12) at lower pressure than the feed stream (3), which is enriched in carbon dioxide compared to the first retentate stream (11), and a second retentate stream (13), which is enriched in methane compared to the first retentate stream (11);(f) withdrawing the second retentate stream (13) from the membrane separation unit (4) as first product stream or further processing the second retentate stream (13), preferably outside of the membrane separation unit (4);(g) recompressing the first permeate stream (10) with the support compressor (9) to obtain a recompressed first permeate stream (14);(h) passing the recompressed first permeate stream (14) to the gas inlet of the permeate stream membrane separation stage (7) and processing the recompressed first permeate stream (14) in the permeate stream membrane separation stage (7) to produce a third permeate stream (15) at lower pressure than the recompressed first permeate stream (14), which is enriched in carbon dioxide compared to the recompressed first permeate stream (14), and a third retentate stream (16), which is enriched in methane compared to the recompressed first permeate stream (14);(i) withdrawing the third permeate stream (15) from the membrane separation unit (4) as second product stream or further processing the third permeate stream (15), preferably outside of the membrane separation unit (4), or sent the third permeate stream (15) to vent;(j) passing the third retentate stream (16) to the gas inlet of the fourth membrane separation stage (8) and processing the third retentate stream (16) in the fourth membrane separation stage (8) to produce a fourth permeate stream (17) at lower pressure than the third retentate stream (16), which is enriched in carbon dioxide compared to the third retentate stream (16),and a fourth retentate stream (18a), enriched in methane compared to the third retentate stream (16);(l) recirculating the fourth permeate stream (17) to the first permeate stream (10), at a third gas recycle point (21) located up-stream of the support compressor (9);(m) recirculating the second permeate stream (12) to the gas stream (1) at a second gas recycle point (20) located up-stream of the main compressor (2) or to the fourth retentate stream (18b) at a fourth gas recycle point located up-stream of the connecting point (19) and downstream of the pressure control unit (22); characterized in that it further comprises the step(k) passing the fourth retentate stream (18a) through a pressure control unit (22), preferably a pressure reduction valve (22), to obtain a fourth retentate stream (18b) at the pressure of the gas stream (1) plus minus 10%, preferably plus minus 5%, more preferred plus minus 2%, even more preferred plus minus 1 %, most preferred at the pressure of the gas stream (1), and recirculating the fourth retentate stream (18b), to the gas stream (1) at a first gas recycle point (19) located up-stream of the main compressor (2);(n) optionally but preferably controlling the pressure on the retentate side of the retentate stream membrane separation stage (6), preferably also of the feed stream membrane separation stage (5), with a pressure control unit (23) located in the second retentate stream (13);(o) optionally but preferably controlling the pressure on the retentate side of the fourth membrane separation stage (8), preferably also on the retentate side of the permeate stream membrane separation stage (7), with a pressure control unit (22) located in the fourth retentate stream (18); and that the main compressor (2) and the support compressor (9) are selected and operated such that the main compressor (2) has a higher compressor load than the support compressor (9), such that ratio of the power consumption of the support compressor (9) to the power consumption of the main compressor (2) CpOwer(9) / CpOwer(2) is in a range of from 0.001 to 0.8, preferably 0.005 to 0.6, more preferred 0.01 to 0.5, even more preferred 0.01 to 0.4, particularly preferred 0.01 to 0.3 and most preferred 0.01 to 0.25; the ratio of the outlet pressure of the support compressor (9) to the inlet pressure of the support compressor (9) psupport-out / psupport-in is in a range of from 1 .05 to 3.5, preferably 1.1 to 3, more preferred 1 .1 to 2.5 and most preferred 1.1 to 2.2; the ratio of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) Pmain-out / Psupport-out * 1 ; the membranes used in separation stages 1 , 2 and 3 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 25 to 100, preferably 30 to 100, more preferred 30 to 90, even more preferred 35 to 90, particular preferred 40 to 80 and most preferred 45 to 70;the membranes used in separation stage 4 having a mixed gas selectivity Sco2 / cH4 at standard pressure and standard temperature of the respective membrane separation unit of 20 to 100, preferably 25 to 90, more preferred 25 to 80 and most preferred 25 to 70.

2. The process according to claim 1 , characterized in that the pressure pPermi of the first permeate stream (10) is in a range of from 1 bara to 1.2 * jp^eed) bara, preferably of from 1 to jp(jeed) bara , and most preferred of 0.95* [p<eed) bara, wherein p<feed) is the pressure of the feed stream (3) and / or the pressure pmain-out at the gas outlet of the main compressor (2) is in a range of from 6 to 40 bara, preferably 6 to 30 bara, more preferred 6 to 20 bara, even more preferred 8 to 16 bara and most preferred 8 to 12 bara; and / or the stage cut SCi of the feed stream membrane separation stage (5) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 1 mol / mol; more preferred 0.08 to 0.95 mol / mol, particularly preferred 0.15 to 0.65 mol / mol and most preferred 0.18 to 0.62 mol / mol and / or the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 0.95 mol / mol; more preferred 0.08 to 0.8 mol / mol, particularly preferred 0.09 to 0.75 and most preferred 0.1 to 0.7 mol / mol and / or the stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 0.99 mol / mol; more preferred 0.08 to 0.95 mol / mol and most preferred 0.1 to 0.9 mol / mol and / or the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.01 to 1 mol / mol, preferably 0.05 to 1 mol / mol; more preferred 0.1 to 0.95 mol / mol and most preferred 0.4 to 0.9 mol / mol.

3. The process according to claim 1 or 2, characterized in that the ratio of the pressure of the first permeate stream (10) to the pressure of the fourth permeate stream (17) pPermi / pPerm4 is in the range of 0.9 to 1 .1 , preferably 0.95 to 1 .05, more preferred 0.98 to 1 .02 most preferred is 1 ; and / or the ratio of the pressure of gas-stream (1) to the pressure of the second permeate stream (12) pgsiI pPerm2 is in the range of 0.9 to 1.1 , preferably 0.95 to 1 .05, more preferred 0.98 to 1 .02 most preferred is 1 ;and / or the ratio of the pressure of the gas-stream (1) to the pressure of the fourth retentate stream (18b) pgsi / pret4 is in the range of 0.98 to 1 .02 most preferably 1 ; and / or that the pressure of fourth retentate stream (18b) to the pressure of the second permeate stream (12) pret4 / pperm2 is in the range of 0.98 to 1 .02, preferably 1 and / or the pressure of the raw gas stream is in a range of from 0 to 1 barg, preferably 0 to 0.5 barg, more preferred 0.01 to 0.4 barg, even more preferred 0.01 to 0.3 barg and most preferred 0.05 to 0.3 barg.

4. The process according to any one of claims 1 to 3, characterized in that the CO2 concentration in the raw gas stream (RGS) Cco2-raw is controlled to be in a range of from 10 to 70 mol%, preferably 20 to 60 mol% and more preferred 30 to 50 mol% and the methane concentration in the raw gas stream (RGS) CcH4-raw is controlled to be in a range of from 30 to 90 mol%, preferably 40 to 80 mol% and more preferred 50 to 70 mol%, wherein the contents of carbon dioxide and methane are selected such from the ranges defined before that both components plus optionally further gases comprised in the raw gas stream (RGS) sum up to 100 mol% of the raw gas stream (RGS) composition, and / or the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 90 to 100 mol%, preferably 95 to 99.9 mol%, more preferred 97 to 99.9 mol%, even more preferred 97 to 99.8 mol%, particularly preferred 97 to 99.5 mol% and most preferred 98 to 99.5 mol% and / or the methane recovery is controlled to be in a range of from 95 to 100 mol%, preferably 96 to 99.99 mol%, more preferred 97 to 99.99 mol%, even more preferred 98 to 99.99 mol%, particularly preferred 99 to 99.99 mol% and most preferred 99.5 to 99.9 mol%5. The process according to any one of claims 1 to 4, characterized in that it is operated at the following process conditions the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, and the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bar, andthe CO2 / CH4 mixed gas selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 25 to 70, preferably 30 to 70, and the ratio pmain-out / psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Varn bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure ppermi of the first permeate stream (10) is in a range of from 1 to Var2h bara, wherein Var2h being the upper limit of the pressure range, and the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from Var4i to Var4h mol / mol, wherein Vam being the lower bound and Var4h being the upper bound of the stage cut range, and the stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.1 to 0.9 mol / mol, and wherein Vam, Vanh, Var2h, Varsi, Varsh, Vam, Vamh, Varsi and Vanh, are regressed variables calculated by use of equation (I)Var = Interceptwith “Intercept” and “ai” being fixed factors taken from the Table 13, xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and xi with i = 4 being the selectivity X4 = Sco2 / cH4of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined in Claim 5 aboveTable 136. The process according to any one of claims 1 to 4, characterized in that it is operated at the following process conditions the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, and the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bar, and the CO2 / CH4 selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 25 to 70, preferably 30 to 70, and the ratio Pmain-out / Psupport-out of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) is in a range of from Varn bara / bara to Vanh bara / bara, wherein Vam being the lower limit of the pressure range and Vanh being the upper limit of the pressure range, and the pressure pPermi of the first permeate stream (10) is in a range of from 1 to Var2h bara, wherein Var2h being the upper limit of the pressure range, and the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC2 of the retentate stream membrane separation stage (6) is in a range of from Var4i to Var4h mol / mol, wherein Vam being the lower bound and Var4h being the upper bound of the stage cut range, and the stage cut SC3 of the permeate stream membrane separation stage (7) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the stage cut SC4 of the fourth membrane separation stage (8) is in a range of from 0.4 to 0.9 mol / mol if the mixed gas membrane selectivity Sco2 / cH4 is 50 or above 50, preferably 50 to 70 and 0.7 to 0.9 if the mixed gas membrane selectivity Sco2 / cH4 is 30, wherein for mixed gas membrane selectivity SCO2 / CH4 of 30 to 50 the lower bound of the SC4 range should be calculated scaling linearly with selectivity between the values 0.7 for SCO2 / CH4 = 30 and 0.4 mol / mol for SCO2 / CH4 = 50 and the upper bound stays at 0.9 mol / mol independent of SCO2 / CH4 , wherein Vam, Vanh, Vamti, Varsi, Varsh, Vam, Vamh, Varsi and Vanh, are regressed variables calculated by use of equation (I)Var = Intercept + xlal(I)with “Intercept” and “ai” being fixed factors taken from the Table 14, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity x4= SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and x4= are values taken from the ranges defined in Claim 6 aboveTable 147. The process according to any one of claims 1 to 4, characterized in that it is operated at the following process conditions the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol%, and the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too, and the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bara, the pressure (psupport-out) at the gas outlet of the support compressor (9) is controlled to be in a range of from pmain-out / Varih to pmain-out / Varn bara, wherein Vam being the lower limit of the range of the pressure ratio pmain-out / psupport-out and Vanh being the upper limit of the pressure ratio range, and the CO2 / CH4 selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 25 to 70, preferably 30 to 70, and the pressure pPermi of the first permeate stream (10) and correspondingly that of the fourth permeate stream (17) is in a range of from 1 bara to Var2h bara, wherein Var2h being the upper limit of the pressure range, andthe stage cut SCi of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, wherein Varsi being the lower bound and Varsh being the upper bound of the stage cut range, and the CO2 concentration in the fourth retentate stream (18) (Cco2(i8>) is controlled to be below an upper bound of Vareh in mole fraction, and the CO2 concentration in the fourth permeate stream (17) (Cco2(i7>) is controlled to be below an upper bound of Var?h in mole fraction, and wherein Varn, Vanh and Var2h, are regressed variables calculated by use of equation (II)Var = Interceptwith “ai” being the coefficients for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 15 below, product pairs or squared terms in equation (II) not specified in Table 15 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined in Claim 7 aboveTable 15and whereinVarsi and Varsh are regressed variables being calculated by use of equation (III)Var = Interceptwith “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 16 below, product pairs or squared terms in equation (III) not specified in Table 16 are zero), Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined in Claim 7 above Table 16 for Varsi and Varshand whereinVareh and Var?h are regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for theproduct of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 17 below, product pairs or squared terms in equation (lll)not specified in Table 17 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined in Claim 7 aboveTable 17 for Vareh and Var?h8. The process according to any one of claims 1 to 7, characterized in that it is operated at the following process conditions the methane concentration CcH4-ret2 in the second retentate stream (13) is controlled to be in a range of from 97 to 99.5 mol% the CO2 concentration in the raw gas stream (RGS) (Cco2-raw) is controlled to be in a range of from 30 to 50 mol%, the rest of the raw gas stream is mainly CH4, other minor components, preferably N2 and / or O2 can be comprised, too,the pressure (pmain-out) at the gas outlet of the main compressor (2) is controlled to be in a range of from 8 to 16 bara, the pressure (psupport) at the gas outlet of the support compressor (9) is controlled to be in a range of from pmain-out / Varih to pmain-out / Varn bara, the CO2 / CH4 selectivity SCO2 / CH4 of the membranes in all four separation stages, at standard conditions of the respective separation stage, is in a range of from 30 to 70 the pressure ppermi of the first permeate stream (10) is in a range of from 1 bara to Var2h bara, the stage cut SC1 of the feed stream membrane separation stage (5) is in a range of from Varsi to Varsh mol / mol, the CO2 concentration in the fourth retentate stream (18) (Cco2(i8>) is controlled to be below an upper bound of Vareh in mole fraction, the CO2 concentration in the fourth permeate stream (17) (Cco2(i7>) is controlled to be below an upper bound of Var?h in mole fraction, wherein Varn, Vanh and Var2h, are regressed variables being calculated by use of equation (II) with “ai” being the coefficients for the single variables, “aij” being interaction coefficients for the product of the given variables, and “a^” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 18 below, product pairs or squared terms in equation (II) not specified in Table 18 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes, and wherein xi , X2 X3 and X4 = are values taken from the ranges defined in Claim 8 aboveTable 18and whereinVarsi and Varsh, being regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “ai»2” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “ai»2” being fixed factors taken from the Table 19 below, product pairs or squared terms in equation (III) not specified in Table 19 are zero), Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined in Claim 8 aboveTable 19 for Varsi and Varshand whereinVareh and Var?h being regressed variables being calculated by use of equation (III) with “ai” being the coefficients tabulated for the single variables, “aij” being interaction coefficients for the product of the given variables, and “ai»2” being coefficients for the squared power of given variables. “Intercept”, “ai”, “aij” and “a^” being fixed factors taken from the Table 20 below, product pairs or squared terms in equation (III) not specified in Table 20 are zero, Xi with i = 1 being the methane content in the second retentate stream (13) xi = CcH4-Ret2, Xi with i = 2 being the pressure at the gas outlet of the main compressor (2) X2 = Pmain-out, Xi with i = 3 being the carbon dioxide content in the raw gas stream (RGS) X3 = Cco2-raw and Xi with i = 4 being the selectivity X4 = SCO2 / CH4 of the membranes and Xi with i = 5 being the outlet pressure of the support compressor (9) X5 = psupport-out, and wherein xi , X2, X3, X4 and X5 = are values taken from the ranges defined in Claim 8 aboveTable 20 for Vareh and Var?h9. A device for separating methane from a gas stream comprising methane and carbon dioxide, comprising:a gas-stream conduit (1) connected to a raw gas source, providing a raw gas comprising methane and carbon dioxide, and to a gas inlet of a main compressor (2) a feed stream conduit (3) connected to a gas outlet of the main compressor (2) and to a feed gas inlet of a feed stream membrane separation stage (5), which is part of a membrane separation unit (4), which further comprises a retentate stream membrane separation stage (6), a permeate stream membrane separation stage (7), a fourth membrane separation stage (8), and a support compressor (9) a first permeate stream conduit (10) connected to a permeate gas outlet of the feed stream membrane separation stage (5) and to a gas inlet of a support compressor (9) a first retentate stream conduit (11) connected to a retentate gas outlet of the feed stream membrane separation stage (5) and to a gas inlet of the retentate stream membrane separation stage (6) a second permeate stream conduit (12) connected to a permeate gas outlet of the retentate stream membrane separation stage (6) and to the gas-stream conduit (1) at a second gas recycle point (20) or connected to a permeate gas outlet of the retentate stream membrane separation stage (6) and to the fourth retentate stream conduit (18b) at a fourth gas recycle point, located upstream of the main compressor (2) and downstream of the pressure control unit (22), a second retentate stream conduit (13) connected to a retentate gas outlet of the retentate stream membrane separation stage (6), the second retentate stream conduit (13) preferably comprises a pressure control unit (23), a recompressed first permeate stream conduit (14) connected to a gas outlet of the support compressor (9) and to a gas inlet of the permeate stream membrane separation stage (7) a third permeate stream conduit (15) connected to a permeate gas outlet of the permeate stream membrane separation stage (7) a third retentate stream conduit (16) connected to a retentate gas outlet of the permeate stream membrane separation stage (7) and to a gas inlet of the fourth membrane separation stage (8) a fourth permeate stream conduit (17) connected to a permeate gas outlet of the fourth membrane separation stage (8) and to the first permeate stream conduit (10) at a third gas recycle point (21) located upstream of the support compressor (9) the first gas recycle point (19) and the second gas recycle point (20), which may be combined to one single gas recycling point, are located upstream of the gas inlet of the main compressor (2) such that the raw gas from the raw gas source, the recycled second permeate stream and the recycled fourth retentate stream are combined to form a gas stream being fed via the gas stream conduit (1) to the gas inlet of the main compressor (2) characterized in that a fourth retentate stream conduit (18a) is connected to a retentate gas outlet of the fourth membrane separation stage (8) and the gas inlet of the pressure control unit (22), a fourth retentate stream conduit (18b) connected the gas outlet of the pressure control unit (22) and to the gas stream conduit (1) at a first gas recycle point (19) the main compressor (2) has a higher compressor load than the support compressor (9) such that the ratio of the power consumption of the support compressor (9) to the powerconsumption of the main compressor (2) CpOwer(9) / CpOwer(2) is in a range of from 0.001 to 0.8, preferably 0.005 to 0.6, more preferred 0.01 to 0.5, even more preferred 0.01 to 0.4, particularly preferred 0.01 to 0.3 and most preferred 0.01 to 0.25; the support compressor (9) is configured such that the ratio of its outlet pressure to its inlet pressure psupport-out / psupport-in is in a range of from 1 .05 to 3.5, preferably 1.1 to 3, more preferred 1 .1 to 2.5 and most preferred 1 .1 to 2.2; the main compressor (2) and the support compressor (9) are configured such that the ratio of the pressure at the gas outlet of the main compressor (2) to the pressure at the gas outlet of the support compressor (9) Pmain-out I Psupport-out>1 the membranes used in separation stages 1 , 2 and 3 are membranes having a higher mixed gas permeability for carbon dioxide than for methane and having a mixed gas selectivity SCO2 / CH4 of 25 to 100, preferably 30 to 100, more preferred 30 to 90, even more preferred 35 to 90, particular preferred 40 to 80 and most preferred 45 to 70, mixed gas permeabilities and mixed gas selectivities are measured at the standard pressure and standard temperature of the respective membrane separation unit; the membranes used in separation stage 4 having a mixed gas selectivity SCO2 / CH4 at standard pressure and standard temperature of the respective membrane separation unit of 20 to 100, preferably 25 to 90, more preferred 25 to 80 and most preferred 25 to 70, mixed gas permeabilities and mixed gas selectivities are measured at the standard pressure and standard temperature of the respective membrane separation unit.

Images