TSA cycle with minimized losses

Abstract

A method for purifying a TSA unit operating according to a cycle comprising an elution-cooling step, said elution-cooling step comprising at least the following two sub-steps: - a first elution-cooling sub-step (C) with, at the adsorber outlet, extraction of a purge gas from the TSA unit as long as the impurity content in said purge gas remains above a predetermined threshold, - a second elution-cooling sub-step (D) using purified product gas, with recycling of the purge gas at the adsorber outlet to at least one other adsorber in adsorption step, the impurity content in said recycled gas then being below the predetermined threshold.

Description

[0001] Description

[0002] Title of the invention: TSA cycle with minimized losses

[0003] The present invention relates to TSA type adsorption gas separation or purification units and in particular to the regeneration of such units. More specifically, it relates to reducing the quantity of gas that it is necessary to extract from the unit to eliminate the impurities that have been removed from the feed gas during the adsorption step. It is therefore intended to increase the extraction yield of the valuable compounds contained in said feed gas.

[0004] Adsorption is widely used to purify or separate gases. Examples include the separation of n- and iso-paraffins, the separation of xylenes and alcohols, the production of nitrogen or oxygen from atmospheric air, CO2 deballasting of combustion gases and blast furnace gases, etc. On the purification side, there are dryers, hydrogen or helium purification, the purification of methane-rich gases such as natural gas or biogas, and the adsorption of trace impurities in many fluids (mercury, NOx, sulfur products, etc.). Processes using adsorption are of several types, depending on whether the adsorbent can be regenerated in situ or not. In the latter case, we speak of “lost charge” adsorption, i.e., to be renewed when the product is saturated with impurities; the term “guard ht” is also sometimes used to describe such purification.When regeneration is done in situ, the adsorbent is said to follow an adsorption cycle which periodically brings it back to a state where it can again begin a separation or purification step.

[0005] Adsorption cycles differ primarily in the way the adsorbent is regenerated. If regeneration is primarily achieved by increasing the temperature, it is a TSA process (Temperature Swing Adsorption). If, on the other hand, regeneration is achieved by decreasing the pressure, it is a PSA process (Pressure Swing Adsorption).

[0006] It should be noted here that regeneration is often carried out by both a heat input and a pressure drop. We will agree here, as is general practice, to call TSA any process for which there is a heat input, whether this input is provided by means of a regeneration gas hotter than the feed gas - let's say at least 10°C - or by any other means such as microwaves, Joule effect, integrated heat exchanger, etc., and this whatever the regeneration pressure.

[0007] In contrast, a PSA unit uses pressure effects exclusively to regenerate the adsorbent, even if internally adsorption and desorption of the constituents cause temperature variations. H2 PSAs and air dryer PSAs are the best known and most characteristic units for this type of operation.

[0008] We are concerned here only with TSA type units according to the preceding definition and in particular with the regeneration of such units. More particularly still, the present invention relates to the reduction of the quantity of gas which must be extracted from the unit to remove impurities from the feed gas.

[0009] Each of the adsorbers implemented in such a unit will therefore include an adsorption period generally at room temperature but possibly at lower temperature, for example at -40°C or at liquid nitrogen temperature for certain cryogenic TSA and a regeneration period including a heating step to desorb the impurities trapped during the previous adsorption cycle. The heating temperature will be for example in a range of 50°C to 300°C for adsorption at room temperature but may be limited to room temperature for cold purification. We can also find in the literature TSAs with hot adsorption, - let's say 300°C - and a regeneration temperature of more than 400°C. We can therefore find TSA type units in a very wide temperature range.

[0010] The regeneration period can include several stages depending on the TSA process chosen. As an example, to illustrate the TSA process, we have chosen one of the best-known purification methods, which has been the subject of numerous publications, namely air purification upstream of a cryogenic air separation unit.

[0011] Typically, a TSA air purification process cycle includes the following steps: a) air purification by adsorption of impurities (water, CO2, traces of hydrocarbons, etc.) at super-atmospheric pressure and room temperature, b) depressurization of the adsorber to atmospheric pressure. c) regeneration of the adsorbent at atmospheric pressure, in particular by waste gases, typically impure nitrogen from an air separation unit and heated to a temperature usually between 100 and 280°C by means of one or more heat exchangers, d) cooling the adsorbent to ambient temperature, in particular by continuing to introduce said waste gas from the air separation unit, but not heated, e) repressurization of the adsorber with purified air from, for example, another adsorber in the production phase or possibly with the air to be purified.

[0012] Typically, air pretreatment devices include two adsorbers, operating alternately, i.e. one adsorber is in the production phase, while the other is in the regeneration phase. The production phase corresponds to the purification of the gas mixture by adsorption of impurities. The regeneration phase includes the depressurization, heating, cooling and repressurization steps mentioned above.

[0013] A step of paralleling the two adsorbers, of varying duration, i.e. from a few seconds to several minutes, is generally added at the beginning or end of the regeneration phase (f). In this case, the cycle can then be represented as in the figure

[0014] I [Fig. 1], the letters (a, b,... f) corresponding to the steps described previously.

[0015] Such a cycle will make it possible to explain what is commonly called the heating stage and the cooling stage.

[0016] It is agreed to call the heating stage, here stage c, the period of the cycle during which a hot regeneration gas is sent into the adsorber or more generally the system is supplied with external energy, by the Joule effect for example.

[0017] The cooling stage, here stage d, is the period of the cycle during which a cold (unheated) gas is sent into the adsorber and where the excess heat is removed, i.e. heat not used for the desorption of the constituents trapped in the adsorbent, if there is excess heat.

[0018] Depending on the TSA process used, heating and cooling can correspond to very different physical situations.

[0019] In fact, these stages are named solely according to the external thermal aspect (supply or not of heat energy to the system) independently of the position of the temperature front and the impurity content within the adsorbent bed.

[0020] To illustrate this point, we can take some extreme cases.

[0021] The heating step carried out by circulating hot gas can be extended until the outlet temperature of the gas inlet is reached, excluding heat losses. All the adsorbent contained is hot and the impurities have been removed.

[0022] If, on the other hand, heating is carried out using electrical resistors embedded in the adsorbent with circulation of a very low gas flow rate intended to facilitate heat transfer, we may end up with a hot adsorbent, impurities desorbed due to the high temperature but still present in the form of interstitial gas in the adsorber.

[0023] In the latter two cases, the system is supplied with the energy necessary both for the desorption of impurities and for the complete heating of the entire adsorbent.

[0024] Conversely, in the most sophisticated systems, the heating is switched off as soon as the amount of heat strictly necessary to desorb the impurities has been introduced into the adsorber. A part of the adsorbent, on the side opposite the hot gas inlet, remains at a low temperature at the end of heating.

[0025] During the so-called cooling step, the regeneration gas, which is now introduced at room temperature, pushes the heat front present in the adsorber and thus helps to heat the adsorbent which is located towards the opposite end.

[0026] If the amount of heat required for desorption was strictly introduced, there was no temperature peak at the outlet. The cooling step actually consisted of pushing the heat front towards the opposite end of the adsorbent tank and purging the impurities.

[0027] In practice, it may be interesting, even in the case of optimization, to introduce during the heating phase, a small excess of heat compared to the quantity strictly necessary to desorb the impurities in order to observe a beginning of increase in temperature at the outlet of the HT guaranteeing a completed regeneration. In this case, the last part of this phase is used almost exclusively for cooling the adsorbent at the inlet of the HT (inlet on the feed gas side, the regeneration being assumed to be done in counter-current). This type of elaborate regeneration which requires controlling a good number of parameters (flow rates, pressures, temperature, quantity of impurities, etc.) to operate correctly is generally only used for large-scale purifications, such as for example on the purifications at the head in front of the air separation units producing more than 1000 tons of oxygen per day.

[0028] An intermediate process consists of interrupting heating as soon as a temperature rise is detected at the adsorber outlet. This generally means that almost all of the impurities have been extracted from the adsorbent. The cooling step then consists of removing the excess heat front.

[0029] From these few examples, it is clear that the impurities can be purged from the adsorber during heating, cooling or partly during each of these two stages. Generally, the impurity content in the regeneration gas passes through a peak or plateau of composition significantly higher than that of the feed gas and then decreases until the end of the stage to reach a residual content which will depend on the entire process. It can be said from now on that in the principle of the invention, we focus more on the impurity profile than on the thermal itself and that we will speak of purge or elution to characterize the moment when the impurities, pushed by a regeneration gas, leave the adsorber. As explained, this moment can also correspond to the conventional heating and / or cooling stage.In the case of air purification, we have seen that a waste product is used for regeneration, which is generally available in large quantities and free of any impurities. Optimizing regeneration will essentially focus in this specific case on energy expenditure and associated investments (heater, etc.).

[0030] If for any reason it is desired to limit the regeneration flow rate, for example because it is only available in limited quantity, or because it is valued elsewhere and therefore has a significant cost or because part of the gas that has just been purified must be used, which is then deducted from the production, in addition to simply optimizing the quantity of heat, there are a number of means known to those skilled in the art.

[0031] Regarding heating, we can cite, but not be exhaustive:

[0032] - Adoption of a higher regeneration temperature, within the limits permitted by the process depending on the adsorbent (risks of aging, deactivation, etc.) and impurities (coking, polymerization, etc.)

[0033] - Intermediate outlet of the regeneration gas with implementation of a second heater, thus allowing with the same circulating fluid to bring more heat to the adsorbent

[0034] - Heater integrated into the adsorbent, Joule effect heating, microwaves... The heat is then supplied in situ, the regeneration gas is then used to transfer it to the adsorbent and to evacuate the impurities.

[0035] - Advanced insulation of adsorbers and pipes (insulation by double gas layer, etc.) Regarding cooling:

[0036] - Optimization in all the cases mentioned above, of heating means to provide just the heat necessary for the desorption of impurities.

[0037] - Intermediate outlet going to a refrigerant allowing the cooled fluid to be used in the second zone of the adsorber.

[0038] As previously stated, impurities are generally desorbed and removed both during the heating stage itself (heater in operation) and during the cooling stage.

[0039] In the general case, i.e. when more heat than strictly necessary is input during the heating phase, the elution part (removal of impurities) ends in practice before the adsorbent is completely cooled. It is possible in a certain number of cases to stop the cooling phase at this point, substantially reducing the need for regeneration gas. A fraction of the adsorbent is still hot and it may be necessary to manage this point in one way or another. A favorable case occurs when during the adsorption phase, the heat front moves faster than the material front, i.e. the impurity front. It is the feed gas itself, purified in the inlet zone, even if it is still partially hot, which cools the still hot adsorbent.There are several variations to implement this idea depending on the operating conditions of the TSA process in question (co-current cooling of the adsorption to preferentially cool the inlet zone), addition of a short partial co-current cooling step of the adsorber inlet zone to facilitate the stopping of impurities during the transition to adsorption. As long as the gas from the adsorber during this adsorption / cooling step complies with the product specification, we generally speak of an adsorption step in this case to characterize this process, the cooling part then being secondary.

[0040] In practice, this means that TSA cycles can be implemented without a real cooling step. This results in a succession of adsorption steps (and automatically associated cooling) and heating (and purging of impurities). Depending on the process downstream of the TSA, it will be necessary to place a refrigerant on the purified gas, which can then exit hot during part of the adsorption step.

[0041] In another vein, adsorbers / exchangers have been proposed in the literature to reduce the amount of regeneration gas. The adsorbent is in immediate contact with a wall which is successively heated and cooled by one or more external fluids. The adsorber can thus be of the tubular exchanger type or a parallel-passage exchanger. The regeneration gas is then only used for local thermal transport and to evacuate the desorbed products (elution effect).

[0042] A majority of the solutions described above to minimize the regeneration gas involve additional equipment (heater, additional refrigerant, etc.) but are above all complex to implement (intermediate outlet and inlet in an adsorber with the corresponding gas distribution problems).

[0043] Other, simpler means have therefore been imagined.

[0044] The most common method is to use the same fluid to first cool a previously heated adsorber, then at the outlet of this adsorber to heat another adsorber which begins its regeneration stage.

[0045] In the case where Ton regenerates with a fraction of the production, we obtain the principle diagram of [Fig. 2], which however requires at least three adsorbers. The feed gas 1 passes through the adsorber 5 and leaves it purified 2. A fraction 3 of the purified gas 2 is used as regeneration gas and first cools the adsorber 7, from which the impurities have previously been extracted, so that it can start an adsorption cycle again, then after being heated in the exchanger 8, this same fraction 3 desorbs the impurities contained in the adsorber 6 and is finally evacuated from the unit as purge gas 4. Thus each adsorber of the unit successively passes through the stages of adsorption, heating and cooling.

[0046] This solution could possibly be combined with the methods mentioned above, for example with a refrigerant and an intermediate heater at the cost of making the unit not shown in the diagram more complex.

[0047] A particularly interesting case is when the impurities are condensable or at least very easy to extract from the regeneration gas because they are present at least temporarily in large quantities in a gas flow rate much lower than that of the feed gas. They are then extracted from the regeneration gas, which is recycled in full into the feed gas or into production depending on the residual impurity content in this gas. There is then no loss of product gas.

[0048] Let us take as an example, to illustrate this process the drying of CO2 whose diagram is represented in [Fig. 3], The feed gas 1 consisting of water-saturated CO2 is dried in a first adsorber 10 until a flow 2 containing only a few ppm of H2O is obtained in order to avoid any risk of freezing in the downstream units, whether it is a simple gaseous CO2 distribution circuit or a liquefier. A fraction 3 of approximately 10% of the dried CO2 2 is used to heat and then cool the second adsorber 11. A refrigerant 12 and a liquid water separator 13 make it possible to obtain a gas which is at most saturated with water. This fluid 9 is injected into the feed gas of the adsorber 10 during the adsorption phase which will therefore have to treat in this way around 110% of the nominal feed gas.The recirculation of the regeneration fluid 9 is ensured in practice by the charge gas compressor 14 which compresses the CO2 from atmospheric pressure to a pressure of the order of 10 bar abs and which must also be sized for 110% of the charge gas flow rate. After compression, a water cooler and a separator pot are generally present (not shown).

[0049] In practice, this type of loop without gas loss, or with minimal loss, can be used as soon as an inexpensive process is available to reduce the impurity level of the regeneration circuit - in which it is momentarily much higher - to a value close to that of the feed gas circuit. This process, which allows a fraction of the impurities to be extracted from the regeneration gas, can consist of cooling / condensation as in the example, in a permeation unit, catalytic reaction or combustion, absorption or even adsorption. It can be seen that in such a unit a flow rate is generally 5 to 20 times lower than the nominal flow rate of feed gas with more favorable conditions concerning the stopping of impurities linked to the respective inlet / outlet contents between the ancillary purification and the TSA purification (for example, from 20 to 1% for the ancillary purification instead of 1% at Ippm for the TSA).

[0050] It should be noted that at equilibrium, that is to say once the TSA material balances are stabilized - which can take a few cycles - all the impurities contained in the charge gas are eliminated by the additional purification.

[0051] A case similar to the one just described corresponds to the case where the unit annexed to the TSA in which the impurities previously contained in the feed gas are stopped makes it possible to directly obtain a gas meeting the purity specification required for production. Such a process can be used, for example, if the sizing of the annex unit depends essentially on the flow rate passing through it and not on the quantity of impurities to be removed. This may be the case for certain washes (absorption) or specific adsorptions.

[0052] [Fig. 4] schematically represents the removal of mercury in the feed gas 1 which is Natural Gas. The mercury is reversibly stopped in the adsorber 10 while a small fraction 3 of the purified gas 2 is used to desorb this constituent from the adsorber 11 during regeneration. The mercury-laden gas 15, after generally having been cooled in a refrigerant not shown, is then completely freed from this impurity, this time irreversibly, in another adsorber 16 whose size is at least an order of magnitude smaller than the same purification which would be placed on the main flow 1 of natural gas. The purified flow 17 can then be mixed with the main purified flow 2. A booster or even a valve (not shown) can be used to ensure the circulation of the regeneration flow.

[0053] It is indeed possible in a certain number of favorable cases, of which we will give an example, to use a simple cycle and to avoid the implementation of a circulator (fan, blower, booster, etc.) as long as it is possible to regenerate with a fraction of the charge gas and that an additional pressure drop between charge gas and production is accepted. Certain Natural Gas (NG) treatments intended to remove part of the water and heavy hydrocarbons, such as benzene or oils, implement such a process.

[0054] The diagrams in Figures 5 and 6 correspond to such a solution. For the record, these configurations can be found in a booklet by GRACE DAVIDSON CHEMICAL from the 1950s, entitled "Natural gas conditioning through adsorption technology" by the authors JM Campbell and WP Cummings. [Fig. 5] is a cycle that can operate with only two adsorbers 20 and 21. The majority of the feed gas 1, approximately 80%, is directly directed to the adsorber in the adsorption stage while the remaining fraction 25 will serve as regeneration gas. During the heating stage, this fraction 25 passes through the exchanger 22 then the adsorber 21 from which it extracts the previously adsorbed impurities. At the outlet, this flow is cooled in the refrigerant 24 from which the impurities are evacuated in the form of the condensate 26. The flow 27 thus deballasted is injected into the feed gas.The purpose of valve 28 is to create the necessary pressure drop on the charge gas to allow the flow of flow 25,27 to circulate. In practice, it is this valve which is used to adjust the regeneration flow rate 25,27. The more this valve is closed, the more the regeneration flow rate increases.

[0055] [Fig. 6] is a cycle operating on the same principle with three adsorbers 20, 21, 22, adsorber 22 being cooled while adsorber 21 is heated. Such an arrangement makes it possible to reduce the fraction of feed gas required for regeneration or to shorten the duration of the stages and thus to use smaller adsorbers.

[0056] It should be noted again that these processes allow regeneration without loss of charge or production gas because we are in the particular case where all impurities are easily condensable. Unfortunately, these latter systems are obviously not appropriate when at least one impurity cannot be easily removed from the regeneration gas. For example, we can assume that it is necessary to remove from a natural gas not only water, heavy hydrocarbons such as C6+ (alkanes, aromatics, etc.) but also H2S or CO2. These latter constituents should be purged from the system using a gas from the process, otherwise they will accumulate in the adsorber and then exit with the production. This will automatically lead to the loss of valuable molecules, in this case methane a priori.

[0057] It should be noted that, in addition to reducing the regeneration flow rate, it may be beneficial to reduce losses of recoverable constituents by removing them from the adsorber that is to be regenerated. There are several techniques known to those skilled in the art to achieve this result, such as saturating the adsorber by deliberately introducing a surplus of impurities after the adsorption-production step (sometimes referred to as rinsing) and recycling the gas produced during this additional step into the unit, or using two adsorbers in series so that only the fully saturated adsorber is regenerated each time. This type of arrangement can be introduced into many processes as a possible improvement but complicates the cycle. The process according to the invention is intended to improve existing solutions when it is desired to reduce the quantity of purge gas that must necessarily be extracted from the TSA unit to remove impurities from the feed gas.These are therefore cases in which, within the purification process, auxiliary purification is not implemented on the regeneration gas to achieve this elimination but on the contrary in which all the impurities to be removed must be found in the purge gas extracted from the TSA.

[0058] It should be noted here that in the event that there are simultaneously some impurities that are easy to extract from the regeneration gas because they are condensable at room temperature, for example, and on the contrary others that are very difficult to separate, it may be interesting to implement two TSA systems in series, the first intended to stop impurities of the first type (for example water and heavy hydrocarbons from NG) and the second the less adsorbable impurities.

[0059] In the case where it is necessary to purge impurities in a flow containing recoverable constituents, one of the means used to reduce the quantity of regeneration gas is to carry out the heating in a closed or quasi-closed loop. [Fig. 7] illustrates such a method. The gas 31 leaving the adsorber 30 during heating is introduced into the compressor or circulator 32 preferably of the fan, blower, centrifugal booster type and after being heated in the exchanger 33 is reinjected at the inlet of the adsorber 30. The compressor 32 is intended to overcome the pressure losses of the circuit and to ensure the adequate flow of gas. A refrigerant 34 can optionally be placed upstream of the compressor to cool the gas leaving the adsorber and bring it back to a temperature compatible with the equipment.The heat extracted at the level of this refrigerant 34 can then be reintroduced into the gas at the outlet of the compressor by means of an economizer exchanger not shown in the diagram. An outlet 35 is generally provided on the loop to eliminate the excess gas caused by heating (desorption, expansion, etc.). A gas inlet 36 (charge gas, purified gas, gas available outside the unit, etc.) is also possibly provided to make, for example, an initial supply of gas. Heating can be carried out at high pressure or after complete or partial depressurization of the adsorber. It is conventional to maintain the pressure of the loop by means of a discharger 37 which makes it possible to automatically extract the excess gas, including a fraction of the impurities. Such heating can be carried out in different sub-stages, for example at increasing temperature levels to avoid coking of certain adsorbed compounds.The gas circulation can be co-current (in the same direction as the adsorption) or counter-current. The adsorbent can be heated completely, i.e. continue heating until the outlet temperature approaches the inlet temperature or, on the contrary, only a determined quantity of heat can be introduced into the adsorber, for example just the heat necessary for the desorption of impurities and the compensation of thermal losses, the outlet temperature then being able to be significantly lower than the inlet temperature. It will be noted that depending on the case chosen, the adsorber can contain the impurities homogeneously between its inlet and its outlet or, on the contrary, that the latter are essentially located in a zone which is still cold at the outlet of said adsorber.It is only the optimization of the complete cycle, with the purging of impurities and the total or partial cooling of the adsorber which makes it possible to retain the best solution according to the requirements of the process.

[0060] Based in part on this principle of loop heating, schemes have been developed in particular by the natural gas processing industry for large decarbonation units. These schemes are probably not all the subject of patents but are generally known to those skilled in the art and constitute, in our view, the Prior Art closest to the solution recommended in this document. The process in question generally consists of linking the steps of adsorption at high pressure 1, depressurization 2, loop heating 3, Purging and Cooling 4 by purified gas with elimination of the outlet gas loaded with impurities until the required quality and the appropriate temperature are obtained and finally repressurization 5 with possibly depressurization gas 5.1 from step 2 and feed gas and / or purified gas 5.2.The number of adsorbers to be implemented and the arrangement of the stages will depend on the technical and economic conditions of the project. Figures 8, 9, 10 show a series of diagrams illustrating, in a non-exhaustive manner, some possibilities for implementing such a diagram: the diagram in [Fig. 8] with 2 adsorbers, that in [Fig. 9] with 4 adsorbers, and finally that in [Fig. 10] according to the same principle as Figure 9 but with 6 adsorbers, 3 of which are adsorption adsorbers, each treating 1 / 3 of the flow rate.

[0061] Cooling being often the sizing point of the Purge and Cooling step, that is to say that the purging of impurities is finished and that the residual content of said impurities would be acceptable while a fraction of the adsorbent is still hot, the previous cycle can be improved in order to reduce the losses of recoverable compounds.

[0062] This last cycle, which is probably the most efficient of the current cycles known to those skilled in the art, will now be described in the context of an example also relating to the treatment of natural gas. It will involve removing the 2 to 3% of CO2 contained in a natural gas to produce a fluid containing only a few tens of ppm. The TSA unit then comprises three adsorbers as shown diagrammatically in [Fig. 11]:

[0063] - The first 40 in the adsorption phase purifies the charge gas 1 and ensures production 2 at the required CO2 specification

[0064] - The second is in heating phase 41 via a recirculation loop of the type described previously

[0065] - The third 42 is in the purge and cooling phase by means of a fraction 3 of the purified gas2 from the adsorber 40 and the gas loaded with impurities 43 is extracted from the unit during this phase until an impurity content compatible with the production purity specification is obtained. From this moment, the gas 44 extracted from the adsorber 42 is returned to production, after possibly being cooled in a refrigerant 45. A valve 46 on the production circuit between the sampling 47 and reinjection 48 points allows adequate circulation of the fluids.

[0066] The basic principle of such a cycle is that the purge and cooling stage is divided into two sub-stages, one essentially intended to eliminate all impurities from the adsorber and the other to complete the cooling.

[0067] This principle can be applied with many variations regarding the number of adsorbers and / or the arrangement of the stages. As in the cycles mentioned above, depressurization and repressurization stages can also be introduced to carry out heating and purging with extraction of the gas at low or medium pressure.

[0068] Such a TSA process with partial recycling of the purge gas in production generally limits the loss of recoverable constituents to approximately 10% of the feed gas flow rate - this percentage obviously being variable depending on the operating conditions - but such a loss, even reduced, may prove difficult to accept in a certain number of cases of high recovery of the feed gas or if the treatment of the purge gas loaded with impurities is very costly because it is subject, for example, to environmental requirements. The high recovery of the feed gas may come from the very nature of the gas (scarcity such as Xenon, helium, etc., production cost such as carbon monoxide or hydrogen using complex units or those consuming energy and / or raw materials) or expensive upstream treatments (purification, transport, etc.).

[0069] In such a case, there is therefore a real interest in developing a TSA type process which minimizes losses during regeneration even more than the cycle described above, which is considered to be the most efficient known to those skilled in the art. More precisely, the problem arises essentially as we have seen, when it is a question of an impurity which is difficult or even impossible to extract, economically at least, from the regeneration gas (non-condensable impurity for example) and which must therefore be purged using a gas from the unit.A solution according to the invention then consists in implementing a method for purifying a feed gas into at least one of its constituents by adsorption using a TSA unit, with production of a purified product gas and extraction of a portion of the purified product gas, called the production stream, from the TSA unit, the TSA unit operating according to a cycle comprising at least one adsorption step at a high pressure of the cycle, a loop heating step using a machine for circulating a heat transfer gas in a loop and an elution-cooling step, characterized in that said elution-cooling step comprises at least the following two sub-steps:.

[0070] - A first elution-cooling sub-step with, at the adsorber outlet, extraction of a purge gas from the TSA unit as long as the impurity content in said purge gas remains above a predetermined threshold,

[0071] - A second sub-stage of elution-cooling using purified product gas, with recycling of the purge gas at the adsorber outlet to at least one other adsorber in the adsorption stage, the impurity content in said recycled gas then being lower than the predetermined threshold.

[0072] According to a variant, the cooling elution step comprises an additional sub-step coming after the first and second sub-steps, said additional sub-step being an elution-cooling by purified product gas, with recycling of the purge gas at the adsorber outlet directly into the production flow, provided that the impurity content obtained after mixing said purge gas with the production flow complies with a given purity specification.

[0073] According to a variant, the elution-cooling step, in particular the first elution-cooling sub-step and the second elution-cooling sub-step, of an adsorber leaving the loop heating step is carried out by purified product gas which has previously passed through another adsorber already eluted and partially cooled during a previous elution-cooling step, in particular in order to complete the cooling of said other adsorber.In particular, the purge gas at the outlet of said adsorber leaving the loop heating step and in the elution-cooling step is extracted from the TSA unit during the first sub-step, the purge gases being recycled during the second sub-step to at least one other adsorber in the adsorption step, as soon as the impurity content of said purge gas becomes lower than the predetermined threshold, and said purge gas is optionally introduced into the production flow as soon as the impurity content obtained after mixing said purge gas with the production flow complies with said purity specification.

[0074] According to a variant, the recycling of the purge gas to at least one other adsorber in the adsorption stage is carried out by means of a booster, said booster possibly being able to be used at another time of the cycle for the circulation of the heat transfer gas in a loop as said machine.

[0075] According to a variant, the cycle further comprises at least one depressurization step, in particular at low or medium pressure, located in the cycle before the first elution-cooling sub-step, the cycle further comprising at least one repressurization step before the second elution-cooling sub-step, the depressurization step being able to take place after the loop heating, before this heating or possibly during the heating.

[0076] According to a variant, the impurities present in the feed gas which may damage an adsorbent of the TSA unit or at least irreversibly reduce its adsorption capacity, impurities such as heavy hydrocarbons, acid compounds or water, are stopped in a pretreatment upstream of the TSA unit, the purge gas extracted from the TSA unit possibly being able to be used for said pretreatment, for example as regeneration gas.

[0077] According to one variant, the feed gas is preferably natural gas or biogas or more generally comprises more than 50% mole of hydrocarbons, mainly methane.

[0078] According to one variant, the charge gas is more than 50 mol% a gas or a mixture of gases belonging to the group formed by hydrogen, carbon monoxide, nitrogen, rare gases, in particular helium, xenon or argon.

[0079] According to a variant, said constituent into which the feed gas is to be purified comprises carbon dioxide and / or hydrogen sulfide.

[0080] The invention also relates to a TSA purification unit comprising a control module for said unit, the control module being configured to implement the method according to one of the preceding claims.

[0081] According to a variant, the content of the gas produced is determined by analysis, preferably before and after the injection point of the purge gas and the result of said analysis is used in the control-command of the TSA cycle, in particular to manage the different elution-cooling stages.According to a variant, the TSA unit comprises two adsorbers in the adsorption stage A1, A2, at least for the duration of the purge gas recycling stage, the two adsorbers then operating respectively at slightly different pressures HP1 and HP2, with HP1>HP2, these pressures being such that the purified gas leaving the adsorber A1 is at a sufficient pressure to compensate for the pressure losses linked to the elution-cooling sub-stage corresponding to said recycling and to be injected at the inlet of the adsorber A2 in order to be purified through it, such circulation of fluids being ensured only by the creation of pressure losses caused by restrictions of the passage section in the pipes, preferably by valves, said pressure losses then being preferably less than 1 bar.

[0082] According to a variant, said pressure losses caused by restrictions in the passage section in the pipes, preferably by valves, are located on the feed gas supply pipes at the inlet of the adsorbers and / or on the purified gas pipes at the outlet of the adsorbers, the cyclical modification of the opening percentages of these valves then making it possible to carry out the selected cycle.

[0083] According to one variant, the gas resulting from the depressurization of an adsorber having completed the adsorption step is used either to partially repressurize an adsorber which will pass into the elution-cooling sub-step by purified product gas with recycling of the gas at the adsorber outlet to the TSA unit, or as elution gas from an adsorber in the elution-cooling sub-step with extraction of the gas at the adsorber outlet from the TSA unit. According to one variant, in addition to possibly gas resulting from depressurization, the gas passing through the adsorber in the elution-cooling sub-step with extraction of the purge gas at the adsorber outlet from the TSA unit is feed gas, purified product gas or a gas outside the TSA unit compatible with the process which will then generally be purged from the circuits before passing to the sub-step with recycling.

[0084] According to one variant, the feed gas from the TSA unit is previously purified, at least partially, of H2O, NH3, alcohols, hydrocarbons higher than C3, mercury, sulfur compounds and / or volatile organic compounds in one or more suitable pretreatments. According to one variant, a stream from the TSA unit which is the subject of the invention is used in at least one of said pretreatments, preferably the stream extracted from the TSA unit loaded with impurities.

[0085] According to a variant, the pressure at which the fraction of the purge gas containing the impurity is extracted from the unit is either freely chosen so as to optimize the process if this fraction is then intended to be sent to the torch, or is imposed by the use made of it, for example this pressure could be at least around 5 bars abs if it is injected into a fuel-gas network.

[0086] The proposed solution does not aim to reduce the quantity of gas used in regeneration itself but to minimize the fraction of this gas that will have to be extracted from the TSA unit as purge gas to eliminate therefrom the impurities contained in the feed gas. To this end, the cycle chosen comprises at least one high-pressure adsorption step, and a regeneration step itself comprising a loop heating step using a machine for circulating the heat transfer gas followed by an elution-cooling step, which comprises at least the following two sub-steps in the order cited below:

[0087] - elution-cooling sub-step with extraction of the purge gas at the adsorber outlet from the TSA unit as long as the impurity content in said gas remains above a predetermined threshold XI,

[0088] - elution-cooling sub-step using purified product gas, with recycling of the gas at the adsorber outlet to the TSA unit, and in particular to an adsorber in the adsorption step, the impurity content in said gas then being lower than the predetermined threshold XI.

[0089] Thus, only the gas relating to the first of these sub-steps will be lost and will determine the extraction yield of the TSA into the valuable compounds.

[0090] It is appropriate here to clarify what is meant by “elution-cooling” in the context of the invention.

[0091] This term corresponds to the step - as well as the two or three sub-steps that compose it - which takes place after the loop heating and which has the dual purpose of eliminating impurities from the adsorber and of cooling the same adsorber sufficiently to allow it to then start an adsorption step again.

[0092] This step therefore consists of passing a gas at ambient temperature through the adsorber, which has just been heated, which will begin cooling the adsorbent. It should be remembered that during cooling, the gas will push the heat front ahead of it and possibly heat the part of the adsorbent located at one end which is still at ambient temperature or at an intermediate temperature. At the same time, the circulating gas will carry with it the impurities which have been desorbed previously, both by piston effect for those in the dead volumes and inter-particle spaces and by elution for those in the pores of the adsorbent. The terms "heating" and "cooling" are therefore relative and if it is necessary to be more precise, it is necessary to define the state of the adsorber, in particular where the temperature front is located.In practice, "elution-cooling" corresponds to the sub-steps during which a preheated adsorber in a loop, totally or partially, will be crossed by a gas flow, with an inlet temperature that can vary over time if necessary, with the dual aim of purging impurities and / or cooling the adsorbent. The purge gas is the flow more or less loaded with impurities and more or less hot which leaves the adsorber in the elution-cooling step. It is this flow which, during the cycle, will be extracted from the TSA unit, then recycled to the inlet of an adsorber in the adsorption step and then possibly recycled into production.

[0093] The novelty of the invention is that a fraction of the purge gas initially taken from production and containing impurities is no longer extracted from the unit but is reintroduced with the impurities it contains into an adsorber of the TSA in the adsorption phase to be itself purified in this same unit.

[0094] A negative consequence of this recycling is that the volume of adsorbent required to stop the impurities will increase. For example, instead of treating a flow rate of 100 of feed gas (standardized flow rate) containing 2% mole of CO2, we will have to treat this same flow rate plus on average a flow rate of 10 containing for example 3% CO2, that is to say 110 at a CO2 content close to 2.1% mole on average.

[0095] Although the adsorber is a little larger and contains more CO2, the amount of gas that Ton will have to extract from the unit and therefore lose is lower in this case than in the basic case. Thus, with a slight oversizing, we will be able to significantly increase the extraction efficiency in the recoverable compounds.

[0096] Simulations of various operating cases confirm the economic gain due to the invention, but it is possible without making complex assessments to understand the favorable mechanisms linked to the invention at stake if you note:

[0097] - On the one hand, the overall quantity of impurities to be extracted from the system remains unchanged despite the recycling of some of them. Indeed, if we imagine an overall material balance of the TSA unit including recycling with only the feed gas as input in this expanded system and the output being the production (purified gas) and the purge gas extracted from the unit, the latter, whatever the case, only contains the impurities brought in by the feed gas. Figure 8, which will be described below, allows this observation to be easily visualized. Everything happens as if the impurities reinjected into the adsorber were going around in circles without any effect on the overall material balance. It should be noted that a few cycles at the start of the unit may be necessary to reach this state of equilibrium.

[0098] - That on the other hand, this same quantity of impurities is extracted from only the initial fraction of the elution gas, a fraction in which it will be found with a content significantly higher than in the basic case where all or almost all of this gas is extracted.

[0099] - The higher the impurity content in the purge gas, the less of this otherwise recoverable gas is needed to extract them.

[0100] The average impurity content in the gas extracted to the outside of the unit can be in an average case 25% higher than it would be without implementing the claimed process. Taking into account the above remarks, the losses of recoverable gas will be reduced by 20%. If the economic conditions justify it, it is possible to go significantly further in reducing losses.

[0101] Thus, contrary to what intuition might suggest, returning CO2 in this way to the inlet of an adsorber in the purification phase leads to a substantial gain in the extraction yield of TSA into recoverable gas.

[0102] The impurity content XI from which the purge gas is recycled at the outlet of an adsorber in the elution-cooling step at the inlet of an adsorber will be the result of an optimization of the TSA unit and recycle assembly. This content can be lower, equal to or higher than the impurity content of the feed gas. It is understood that there is an optimum between not recycling anything (base case) and recycling too much which will lead to a significant increase in the volume of adsorbent required and therefore which will make heating and cooling longer due to the increase in the mass of adsorbent. Knowing the process to be implemented described in this document, the person skilled in the art now has the tools to find the best solution according to the techno-economic conditions of the project.

[0103] [Fig. 12] shows these characteristic stages very schematically.

[0104] It shows in particular the different stages or sub-stages through which an adsorber of the TSA unit passes without prejudging the number of adsorbers that this unit comprises. In order, an adsorption stage A at high pressure HP with purification of the feed gas 1 and production of a purified flow 2, a loop heating stage B, an elution-cooling sub-stage C with extraction of the purge gas 55 from the unit 57 as long as the content of impurities in it remains higher than a predetermined threshold XI and finally an elution-cooling sub-stage D by a fraction of the production 3 and recycling of the purge gas 56 from the adsorber to the inlet of an adsorber in adsorption stage A. Generally speaking, the last three stages B, C, D correspond to regeneration.In the text, we generally speak of sub-steps with regard to states C and D, because in both cases it is a question of cooling an adsorber and simultaneously extracting impurities from it. If we do not particularly seek to distinguish these two sub-steps, we call step C, D the overall step. In the Figures or descriptions, we sometimes identify an adsorber in a given step by its numbering in the cycle (adsorber 60, 61 . . . for example) and by the type of step in which it is then found (for example B if it is loop heating). The number of adsorbers involved, as indicated previously, is not a characteristic of the process but corresponds to the best arrangement that the person skilled in the art can make according to the constraints of the site and the economic conditions.Thus in Figure 12, 4 adsorbers 50, 51, 52 and 53 are shown, one in each of the stages A, B, C, D mentioned, but there could just as well be several adsorbers in adsorption stage A while stages C and D could follow one another on the same adsorber. It is indeed the presence of the main stages which characterize the proposed process, namely therefore a regeneration comprising heating in loop B, followed first by an elution-cooling stage C with extraction from the unit of the purge gas at the adsorber outlet then by an elution-cooling stage D with the purified gas and with recycling of the purge gas at the inlet of an adsorber in adsorption stage A.

[0105] It should be noted that until now we have spoken of heating and then cooling to ambient temperature. These temperature levels do not constitute a limitation to the scope of the invention, which may be applied to cold or cryogenic purification, for example with adsorption at the temperature of liquid nitrogen and with heating to ambient temperature, or to purification operating between two relatively high temperatures, for example 100°C in adsorption and 350°C in regeneration.

[0106] As with the majority of cycles used in adsorption, there are a number of variations based on this basic principle which allow, at least in certain cases, performance to be improved.

[0107] In particular, the elution-cooling step may comprise an additional sub-step coming after the two sub-steps described previously, said additional sub-step being an elution-cooling also by purified product gas, with recycling of the purge gas at the adsorber outlet in the production as soon as the impurity content in said gas is lower than a second predetermined threshold X2, with X2 <X1. C’est le cas si l’élution des impuretés est plus rapide que le refroidissement de l’adsorbant, autrement dit si le front de matière est plus rapide que le front thermique. On va se retrouver en cours d’étape à un moment où la teneur résiduelle en impuretés va être compatible avec la spécification de la production. Il est alors possible d’injecter directement le gaz de purge en sortie d’un adsorbeur en étape d’élution-refroidissement dans ladite production. L’exemple suivant va permettre d’expliciter ce point.A feed gas containing 2 mol% CO2 is purified of CO2 to obtain a production whose CO2 content must be less than or equal to 50 ppm mole. A standardized flow rate of 100 is treated and a flow rate of 10 is used for the elution-cooling step with recycling. It is assumed that during the recycling period, the purified gas (production) leaves with 20 ppm CO2, i.e., it is below the specification of 50 ppm. Such a margin is common because it is generally only at the end of the adsorption phase that the limit value is approached. The recycling step is started by injecting a gas containing 3.5% CO2 into the adsorber inlet, a value chosen as optimal for the process according to the invention in this specific case.From the moment the impurity content reaches 320 ppm in the recycled gas, this gas can be injected directly into the production which will then consist of a flow of 90 to 20 ppm of CO2 and a flow of 10 to 320 ppm, i.e. the equivalent of a flow of 100 to 50 ppm of CO2.

[0108] In this case, the TSA cycle will include 3 elution-cooling sub-steps, following the definition given previously:

[0109] - a step with extraction of the purge gas loaded with CO2 to the outside of the unit, a step using production gas as elution-cooling gas, with recycling of the purge gas at the outlet of an adsorber in the elution-cooling step to the inlet of an adsorber in the adsorption phase with a CO2 content ranging from 3.5% mol to 320 ppm, these two steps being characteristic of the present invention,

[0110] - a stage where this same gas with a content of less than 320 ppm CO2 is injected into the production downstream of the elution gas sampling point.

[0111] The final step in this process should be considered as an improvement of the principle of the invention. It allows a slight reduction in the volume of adsorbent required but requires additional regulation based, for example, on analyses and flow rates. It will generally not be used systematically.

[0112] Here we have taken the example of CO2 as an impurity. In practice, it can be a well-identified constituent such as CO2, H2S, CH4, etc., or a set of constituents characterized as being part of a group (CO2, H2S, COS, etc. or CO, CH4) or by a physical property (for example, solidification point, total sulfur content, etc.).

[0113] It is easy to generalize to n impurities and recycle the regeneration gas only when each content Zi is less than or equal to the value of the corresponding specification Zli in the produced gas. On the contrary, it may be the overall sum of the impurities that matters. The threshold X2 on which the transfer of recycling from the inlet of an adsorber to production depends can be determined experimentally on site based on analyses or previously by simulation and will remain the same as long as the operating conditions and the state of the adsorbent do not vary. Otherwise, a threshold X2 can be planned as a function of a few suitable parameters (feed gas flow rate, impurity content, feed gas inlet temperature, etc.). Simulation software can be used to test many potential operating cases and derive the appropriate X2 function. Such software can also operate in parallel with the unit and adapt X2 to the cycles in progress.It will generally be appropriate to provide for control of the production content with, if necessary, corrective actions relating to the cycle.

[0114] Still within the scope of the invention, a variant makes it possible, at least in certain cases, to further enrich the fraction of gas that will be extracted from the unit with impurities and therefore, with the same reasoning as previously, to thereby reduce the loss of recoverable gas. In this other variant, the elution-cooling step of an adsorber leaving the loop heating step is carried out by purified product gas which has previously passed through an adsorber which has already been at least partially eluted and partially cooled during a said elution-cooling step carried out previously in order in particular to complete its cooling. The purge gas leaving the adsorber located in the present elution-cooling step is extracted from the TSA unit during a first sub-step, then recycled to the TSA unit, the recycling being carried out in particular to an adsorber in the adsorption step as in the basic solution.

[0115] [Fig. 13] shows such a sequence with respectively the adsorbers 60 in loop heating step B, 61 which successively passes through the two elution-cooling sub-steps C and D with extraction of the purge gas loaded with impurity 63 then with recycling of this gas 64 to an adsorber in adsorption step A not shown, while the adsorber 62 which, after having previously undergone a step equivalent (C then D) to that in progress with the adsorber 61, finishes its cooling and the elution of the residual impurities during a step E. The elution gas, whatever its origin, preferably here a small fraction 3 of the purified gas, first passes through the adsorber 62 then the adsorber 61.It is understood that in this way, the flow rate of elution-cooling gas used twice, to finish the cooling (step E) of the adsorber 62 and to evacuate the impurities from the adsorber 61 (steps C then D), is lower than the flow rate required in the basic case and therefore the impurity content in the gas coming from the adsorber 61 is on average significantly higher than in the basic case. With the same explanation as above, this leads to needing a smaller quantity of gas to be extracted to eliminate the impurities from the system. One could determine the relationship between the quantity of gas to be extracted from the unit and the optimum number of adsorbers to be swept in series, let's say three or four, but it seems that only special cases could lead to going beyond two adsorbers in series. Note that an exchanger - not shown in Figure 13 - may be necessary to cool or heat the gas between adsorbers 62 and 61.

[0116] Concerning the recycling of the purge gas at the outlet of an adsorber in the elution-cooling stage, in a first variant this gas is compressed in order to compensate for the various pressure drops and to be able to be introduced at the inlet of an adsorber in the adsorption phase. The overpressure created by this machine must compensate in particular for the pressure drop between the inlet and outlet of the adsorber in the purification stage and the various pressure drops linked to the elution-cooling stage (adsorber, piping, exchanger, etc.). The booster can be placed anywhere on the regeneration circuit, upstream or downstream of the adsorber (or adsorbers) in the elution-cooling stage.

[0117] It should be noted that if the cycle arrangement allows it, this booster can be the machine allowing the heating in a loop. In this case, it is possible to invest in only one compressor but this entails additional constraints (compatible durations of the stages, characteristics of the machine which must correspond to a double use, etc.). If these constraints are acceptable, this is obviously an interesting solution.

[0118] A variant that will often be preferred, in particular when dealing with large gas flow rates that can justify the use of multiple adsorbers in the adsorption phase, consists of carrying out the recycling without a machine. For this, the TSA unit comprises two adsorbers in the adsorption stage A1 and A2, at least for the duration of the elution gas recycling - cooling stage, the two adsorbers then operating respectively at slightly different pressures HP1 and HP2, with HP1>HP2. These pressures HP1 and HP2 are such that the purified gas leaving the adsorber A1 is at a sufficient pressure to compensate for the pressure losses linked to the elution-cooling sub-stage corresponding to said recycling and to be injected at the inlet of the adsorber A2 at pressure HP2 in order to be purified through it.Thus, the necessary circulation of fluids is only ensured by the creation of pressure losses caused by restrictions on the passage section in the pipes, preferably by valves, said pressure losses then being preferably less than 1 bar.

[0119] Said pressure losses caused by restrictions in the passage section in the pipes, preferably by valves, are located on the feed gas supply pipes at the adsorber inlet and / or on the pipes at the adsorber outlet of the purified gas before the general production manifold, the cyclical modification of the opening percentages of these valves then making it possible to carry out the selected cycle.

[0120] For example, at least for the duration of the purge gas recycling step, two adsorption adsorbers operate at pressures equal to 40 bar abs and 39 bar abs respectively. Such a pressure difference is sufficient to ensure the circulation of all the fluids involved in the process according to the invention.

[0121] The gas intended to carry out the elution will be taken from the outlet of the adsorber at the highest pressure, i.e. here at 39.5 bar abs, will elute the regenerating adsorber, will pass through any necessary exchangers and will then be available at 39.1 bar abs and can be injected at the inlet of the second adsorber. In such a case, the purified gas will be available at around 38.5 bar abs. It is understood that by adjusting the dimensions, it is possible to carry out this process with significantly lower pressure losses, in particular significantly less than 1 bar.

[0122] There are several ways to manage recycling based on this principle in practice. Figures 14, 15 and 16 represent three different cases that can be implemented. In all cases, the adsorber 70 in adsorption step A1 is at the highest pressure HP1, while the adsorber 71 in adsorption step A2 is at a slightly lower pressure HP2. These diagrams only show the adsorbers in the adsorption phase necessary to understand this aspect of the process and an adsorber 72 in the elution-cooling sub-step with recycling D. The TSA unit obviously includes other steps such as loop heating and the elution-cooling sub-steps with extraction of the purge gas loaded with impurities from the unit, steps that are not useful for explaining the recycling principle.

[0123] In [Fig. 14], the adsorber 70 only purifies the feed gas flow 75 necessary to provide the elution-cooling gas 73 while the adsorber 71 treats the complementary feed gas flow 76 and the recycled gas 74.

[0124] In [Fig. 15], adsorber 70 purifies the entire flow of feed gas 1 while adsorber 71 only treats the recycled flow 74.

[0125] Finally, in [Fig. 16], the adsorber 70 purifies only a fraction of the feed gas 75 so as to produce a purified gas flow rate greater than the requirement 73 of the elution-cooling sub-step with recycling while the adsorber 71 treats the additional feed gas flow rate 76 and the recycled gas flow rate 74. In the latter case, said fraction of the feed gas treated by the adsorber 70 is preferably between 30 and 70% of the feed gas flow rate in order to at least partially balance the flow rates.

[0126] As shown in these figures, the circulation of the different fluids (feed gas, elution flow, purified gas, etc.) is ensured by the creation of pressure losses caused by flow restrictions, preferably valves, pressure losses preferably less than 1 bar. These are valves 77 (figure 14), 78 (figure 15), 79 and 80 (figure 16). Only the valves useful in the configuration of the three diagrams have been shown. As this is a cyclic process in which each adsorber passes through each of the stages, said valves are in practice present on each of these adsorbers. The same applies to the pipes, of which only those in service appear on the diagrams.

[0127] It should also be noted that this recycling method consisting of taking a purified flow leaving a first adsorber Al in the adsorption stage, using it in the cycle and then injecting the gas flow resulting from said use into another adsorber A2, also in the adsorption stage but at a lower pressure than Al, is not limited to the use described above but can find other applications.

[0128] A given adsorber will generally follow steps 70 and 71 successively in an order to be determined according to the arrangement chosen. For example, in the case of Figure 15, the adsorber 70 only processes a low flow rate of feed gas during this sub-step corresponding just to the necessary elution-cooling flow rate while during the same time the adsorber 71 purifies most of the feed gas and the recycled gas. The adsorber 70 in the Al state, much less loaded with impurities, can then move to the A2 state. Conversely, the low flow rate passing through the adsorber 70 can make it possible to stiffen the material fronts and better use the adsorbent close to saturation. Here again, simulation software for cyclic adsorption processes can allow fine optimization of the process. On site, analyzers will ensure the proper operation of the unit, in particular the purity of the production.The means for regulating the cycle, in particular adapting the cycle time to operating conditions and thereby determining the duration of the different stages and internal flow rates, are not described in this document because they are known and common to many TSA units.

[0129] It should also be noted that the process has been described with two adsorbers in the adsorption phase A1 and A2, respectively adsorbers 70 and 71, but it is clear that any number of adsorbers equal to or greater than 2 can be used within the scope of the invention. As already explained previously, the step of recycling the purge gas at the inlet of an adsorber can be followed by a step during which this same gas is injected directly into the purified gas as soon as the gas obtained after mixing complies with the production purity specification. It should be noted in this regard that the impurity content in the recycled gas decreases while the content in the purified gas varies in the opposite direction. This can make it possible to recycle a relatively polluted gas to production.In the previous example, if the CO2 has not yet penetrated into the production (content 0 ppm), recycling can begin directly into the product as soon as the CO2 content falls below 500 ppm to obtain production at 50 ppm of impurity. Such a solution will need to be validated experimentally on site and / or using cyclic adsorption process simulation software. Indeed, this depends on the respective kinetics of impurity breakthrough in adsorption and the decrease in these contents in elution. As a result, an impurity peak cannot be totally excluded a priori. However, there are ways known to those skilled in the art to smooth out such a peak if it were to occur, such as adding a small volume of adsorbent to the production circuit.

[0130] It should be noted that the elution gas flow rates may be different depending on the different elution-cooling sub-stages (extraction of impurities, recycling at the inlet of an adsorber, recycling to production) and even during the same sub-stage, given that a majority of valves are control valves and not TOR (On / Off) valves and that the machine(s) are equipped with means to control the compressed flow (speed variator, throttling or bypass valve). For example, the person skilled in the art may adopt speed conditions favoring a piston effect to expel the gas heavily laden with impurities found in the adsorber having completed the loop heating stage by limiting the mixing effects (diffusion, dispersion, etc.) or conditions favoring the desorption of residual impurities at the end of the stage. Such choices will also depend on the pressures retained during these stages.

[0131] According to a variant, the cycle further comprises at least one low or medium pressure depressurization step before the elution-cooling step with extraction of the purge gas and at least one repressurization step before the elution-cooling step with recycling. This depressurization step can be carried out after the loop heating, before this heating or possibly during heating.

[0132] Indeed, for the same gas flow rate, the elution effect will be more effective if the pressure is reduced during this stage because the actual volume of circulating gas is then increased and desorption is facilitated. On the other hand, it will then be necessary to repressurize the adsorber and carry out the end of the purge and cooling at high pressure to more easily recycle the purge gas at the inlet of an adsorber in the adsorption stage, therefore at HP. It should be noted that for cooling, the pressure is secondary because it is no longer the volume flow rate that counts but the molar flow rate.

[0133] This depressurization step can be done after the loop heating, before this heating or possibly during heating. Each of these solutions can have advantages and / or disadvantages depending on the characteristics of the project. High pressure heating will require a gas recirculation machine sized for this high pressure but of reduced size. This will generally be a blower type machine. Conversely, low pressure heating will require a machine of significantly larger size but of different technology, possibly simpler to implement, such as a fan. For the same circulating flow rate, the pressure drops across the adsorber will be higher at low pressure and will lead to higher energy expenditure. On the other hand, at HP, these same pressure drops may be too low and lead to risks of poor gas distribution.The Professional will decide on the best solution on a case-by-case basis.

[0134] In the event that there is depressurization, particularly before the heating stage, the gas resulting from this depressurization can have different uses.

[0135] It can be directly evacuated from the unit, thus participating in the purging of impurities. In this case, it will preferably be a counter-current depressurization, i.e. in the opposite direction to the direction of circulation of the gas during adsorption. The interest of such a depressurization will have to be evaluated according to the impurity content in the gas thus extracted from the unit compared to the corresponding content during the specific elution-cooling step with extraction of impurities in the purge gas. Too low a content in the depressurization gas corresponds to a loss of recoverable product and such a step will then have to be avoided.

[0136] According to another variant, the gas resulting from the depressurization of an adsorber having completed the adsorption step is used either to partially repressurize an adsorber which will pass into the elution-cooling sub-step with recycling of the gas at the adsorber outlet to the TSA unit, or as elution gas from an adsorber in the elution-cooling sub-step with extraction of the purge gas at the adsorber outlet from the TSA unit. If the elution gas during the recycling sub-step D is always purified product gas, the gas passing through the adsorber in the elution-cooling sub-step C with extraction of the gas at the adsorber outlet from the TSA unit can come from different sources.This may be, in addition to possibly the gas from a depressurization as indicated above, the feed gas, the purified product gas or even a gas outside the TSA unit compatible with the process, for example nitrogen, a gas which will then generally be purged from the circuits before moving on to the sub-stage with recycling.

[0137] We have seen previously that repressurization could possibly be carried out in part with depressurization gas. In any case, it is appropriate to use another pressurized gas to reach the desired pressure, close to the adsorption pressure. We will then use the feed gas and / or the purified gas for this purpose. We recall here that the elution-cooling step with recycling of the gas at the level of an adsorber in the adsorption step is carried out with purified gas and that it is one of the characteristic steps of the invention.

[0138] Before being purified in the TSA unit according to the invention, one or more other impurities may have been removed from the feed gas, by means of a suitable process (filtration, washing, catalysis, etc.), in particular by means of another adsorption process. These may be compounds in the form of traces that are easy to adsorb but very difficult to desorb. In this case, a guard tank will generally be used, the adsorbent of which can be changed or regenerated after several weeks or even several months. The adsorbent will then be, for example, activated carbon (physical adsorption) or a doped adsorbent (chemisorption), the constituents thus stopped being able to be traces of benzene, naphthalene, oil, mercury, sulfur compounds and / or volatile organic compounds.Generally speaking, these will be products which we do not want to pollute the adsorbent intended to stop the main impurities in the TSA unit according to the invention, in particular during the loop heating stage.

[0139] According to a variant, the feed gas of the TSA unit is first purified, at least partially, of H2O, NH3, alcohols, intermediate hydrocarbons (C6-C8), regenerable compounds but which can significantly reduce the adsorption capacity of an adsorbent even in trace form, for example a few tens of ppm. In the case of easily condensable or very soluble constituents, a first TSA will be used for this purpose and these impurities are extracted during regeneration in liquid form at a two-phase separator after cooling, and the gas leaving said separator is recycled to the inlet of an adsorber. Such a process is known and has been briefly described above among the potential solutions for avoiding losses of recoverable constituents.

[0140] According to a variant, a flow from the TSA unit which is the subject of the invention is used in at least one of the pretreatments cited above, preferably the purge gas extracted from the TSA unit during step C and loaded with impurity. According to a variant, the pressure at which the fraction of purge gas containing the impurity can be extracted from the unit during step C is either freely chosen so as to optimize the process according to possible process constraints (gas velocity, attrition, pressure drops, etc.) if this fraction is then sent to the torch, or is imposed by the use made of it, for example this minimum pressure could be around 5 bar abs if it is injected into a fuel-gas network.

[0141] In this way, various couplings can be imagined between the upstream unit, which will therefore be able to use at least a fraction of the purge gas extracted from the TSA, and the downstream unit, which may be, for example, a cold box which can provide a gas flow used during heating or the first elution-cooling step according to the invention without departing from the characteristics of the present invention. This may also involve thermal coupling.

[0142] Such a process will be applied in particular to the purification of natural gas or biogas, whether to achieve a purity specification linked to a network or to allow downstream treatment, for example liquefaction. The feed gas is then generally available between 10 and 50 bar abs, at ambient temperature and the impurities to be stopped at this level are essentially CO2 and / or H2S. The loss of natural gas or biogas linked to regeneration can then be only a few %, from 3 to 6% for example by applying the process according to the invention.

[0143] Another area of ​​application for the process will be charge gases whose cost is high due to their rarity or the production process. This category includes, but is not limited to, hydrogen, carbon monoxide, but especially argon, helium, krypton, xenon and special gas mixtures intended for electronics, for example.

[0144] The purification of a helium flow into CO2 is taken as an example to explain the characteristics specific to the cycle according to the invention in a more general framework which also includes one or more additional steps, which does not modify the scope of said invention.

[0145] More specifically, we want to continuously purify a flow of 250 Nm3 / h of helium polluted by water, CO2 and other light compounds that will be treated cryogenically and that do not intervene at our level. We will now refer to helium as the feed gas to be purified. We are looking to obtain a fluid with less than one ppm of residual impurities, preferably with around ten ppb in water and CO2. Given the cost of helium, we are looking for the maximum yield of this gas. A PSA unit, for example, makes it possible to achieve the purity specification but with losses of 10 to 20% depending on the complexity of said PSA. In this case, the recompression of the waste for a second treatment that would improve the yield complicates the process and leads to additional pollution risks.

[0146] The solution according to the invention will consist here in first using a first TSA in a conventional manner to completely dry the gas. The regeneration flow is entirely recycled, the water being extracted by condensation before introducing this flow into the adsorber in the adsorption phase. Since helium is not soluble in water, the loss can be considered zero. This is, as already mentioned, a known system, of the type shown in Figure 2 for example.

[0147] The problem now is to remove the 5000 ppm of CO2 from the dried helium stream with the best possible efficiency.

[0148] A first possible cycle according to the invention consists of using a TSA unit with 2 adsorbers and a circulator whose purpose is to compensate for pressure losses and allow the circulation of fluids between the inlet and outlet of the adsorbers. This is then a dry, non-lubricated and sealed machine allowing the fluid passing through it to be overpressurized by around a hundred millibars. [Fig. 17] represents the two adsorbers 100 and 200 with the different stages they follow during the cycle.

[0149] The description of the cycle begins when the adsorber 100, which has been previously regenerated, enters adsorption phase A and the adsorber 200, which has finished its adsorption phase, enters regeneration. This regeneration phase includes steps 201 to 204 characterized by their function:

[0150] - Step A3 (reference 201) of saturation of the adsorbent ht with CO2 by passing a fraction 210 of the flow of feed gas 1. The gas produced which contains more and more CO2 is recycled to the inlet of the adsorber 100 by means of the circulator 205. This step consists of expelling the maximum amount of helium from the mass transfer zone of the adsorbent before moving on to heating and recycling this helium in the process.

[0151] It should be noted that there are other ways of saturating the adsorbent, such as injecting CO2 to displace the helium, it being understood that this additional CO2 must also be evacuated from the system. It would also be possible to carry out a partial co-current depressurization, but in this case it is advisable to have a third depressurized adsorber to collect this gas. This step A3 and the variants cited below constitute in this case an improvement of the process according to the invention but are not one of its characteristics, unlike the following steps:

[0152] - Stage B (reference 202) of closed loop heating during which the initially cold gas is compressed in the circulator 206, heated to 200°C in a heater 207 and introduced into the adsorber. The CO2 is gradually desorbed, the pressure tends to rise in the loop due to the CO2 released and the increase in temperature. The CO2 content in the loop, initially less than 1%, increases to reach a few tens of %. It may be advisable during this stage to purge gas from the circuit to maintain the pressure at the desired level. This purge is not shown here.

[0153] - Step C (reference 203) of purging the CO2 which is carried out by introducing purified helium 211 coming from the adsorber 100 into the adsorber 200. This helium very rich in CO2 is evacuated from the unit until the CO2 content in this gas is of the order of one or a few %.

[0154] - Step D (reference 204) of end of CO2 purge and cooling of adsorber 200 by introduction of purified helium 212, with recycling of this gas to the inlet of adsorber 100 by means of circulator 208 until the target temperature is reached.

[0155] During the same period of time (one phase of the cycle), the adsorber 100 is in adsorption step A at the high pressure of the cycle. It passes through sub-steps 101, 102, 103 and 104 corresponding respectively to steps 201, 202, 203 and 204 of the adsorber 200.

[0156] We then start a new cycle by reversing the role of adsorbers 100 and 200. It should be noted that the boosters marked 205, 206 and 208 are not in operation simultaneously. It may then be the same machine equipped with a system capable of adapting the compressed flow rate to the needs (variable speed, suction throttle valve, bypass valve, etc.). Such a machine could also boost the purge flow extracted from the unit during step 203 if necessary.

[0157] The process chosen here is relatively simple because the flow rate to be treated is low and there are few disadvantages to oversizing the adsorber. The cycle time in particular can be greater than 8 hours while remaining with small equipment compared to the majority of equivalent industrial units (natural gas flow rate more than 100 times higher). The regeneration temperature can be high because, in the absence of humidity, there is no degradation of the adsorbent, in this case the zeolite. A temperature of 250°C or more can be considered to practically eliminate all CO2 adsorption, even at high partial pressure. Elution will essentially consist of pushing the CO2 out of the adsorber. The CO2 content at cutoff, i.e. at the moment when recycling begins, will be essentially determined by simulation, by scanning different values.On-site adjustment is possible by simultaneously adjusting the cycle time to adapt the sizing to the quantity of impurity to be stopped. The geometry of the adsorbers and connecting pipes must also be optimized to limit dead volumes that are full of helium. In the same vein, dense filling of the adsorber will also be beneficial. The helium loss can be very low under these conditions, in the order of a few %.

[0158] It could be further reduced if a small permeator were placed on the fraction extracted from the TSA and if a low-pressure fluid rich in helium were recovered, but this is not within the scope of the invention.

[0159] The invention generally concerns all TSA-type units implementing the process according to the invention, regardless of the recoverable constituents, the impurities to be extracted, the adsorbents used, the respective adsorption and regeneration pressures and temperature levels, the number of adsorbers used and any additional steps.

Claims

Claims

1. A method for purifying a feed gas into at least one of its constituents by adsorption using a TSA unit, with production of a purified product gas and extraction of a portion of the purified product gas, called a production stream, from the TSA unit, the TSA unit operating according to a cycle comprising at least one adsorption step (A) at a high pressure of the cycle, a loop heating step (B) using a machine for circulating a heat transfer gas in a loop and an elution-cooling step, characterized in that said elution-cooling step comprises at least the following two sub-steps: - a first elution-cooling sub-step (C) with, at the adsorber outlet, extraction of a purge gas from the TSA unit as long as the impurity content in said purge gas remains above a predetermined threshold, - a second elution-cooling sub-step (D) using purified product gas, with recycling of the purge gas at the adsorber outlet to at least one other adsorber in the adsorption step, the impurity content in said recycled gas then being lower than the predetermined threshold.

2. Method according to claim 1, characterized in that the cooling elution step comprises an additional sub-step coming after the first and second sub-steps, said additional sub-step being an elution-cooling by purified product gas, with recycling of the purge gas at the adsorber outlet directly into the production flow, provided that the impurity content obtained after mixing said purge gas with the production flow complies with a given purity specification.

3. Method according to one of the preceding claims, characterized in that the elution-cooling step (C, D) of an adsorber leaving the loop heating step (B) is carried out by purified product gas which has previously passed through another adsorber (E) already eluted and partially cooled during a previous elution-cooling step (C, D).

4. Method according to one of the preceding claims, characterized in that the recycling of the purge gas to at least one other adsorber in the adsorption stage is carried out by means of a booster, said booster possibly being able to be used at another time of the cycle for the circulation of the heat transfer gas in a loop as said machine.

5. Method according to one of the preceding claims, characterized in that the cycle further comprises at least one depressurization step, in particular at low or medium pressure, located in the cycle before the first elution-cooling sub-step, the cycle further comprising at least one repressurization step before the second elution-cooling sub-step, the depressurization step being able to be carried out after the loop heating, before this heating or possibly during the heating.

6. Method according to one of the preceding claims, characterized in that the impurities present in the feed gas which may damage an adsorbent of the TSA unit or at least irreversibly reduce its adsorption capacity, impurities such as heavy hydrocarbons, acid compounds or water, are stopped in a pretreatment upstream of the TSA unit, the purge gas extracted from the TSA unit possibly being able to be used for said pretreatment, for example as regeneration gas.

7. Method according to one of the preceding claims, characterized in that the feed gas is preferably natural gas or biogas or more generally comprises more than 50 mol% of hydrocarbons, in particular mainly methane.

8. Method according to one of claims 1 to 6, characterized in that the feed gas is at more than 50 mol% a gas or a mixture of gases belonging to the group formed by hydrogen, carbon monoxide, nitrogen, rare gases, in particular helium, xenon or argon.

9. Method according to one of the preceding claims, characterized in that said constituent into which the feed gas is to be purified comprises carbon dioxide and / or hydrogen sulfide.

10. TSA purification unit comprising a control module for said unit, the control module being configured to implement the method according to one of the preceding claims.

Images