METHOD FOR MANAGING A FAULT IN A PRESSURE SWITCHING ABSORPTION GAS TREATMENT UNIT

Description

[0001] The present invention relates to a method for managing a pressure swing adsorption (PSA) gas processing unit, particularly for producing an enriched gas stream from a feed gas stream, and more specifically to pressure swing adsorption processing units (PSA units) employing at least six adsorbers. It also relates to a unit operating according to this method.

[0002] The invention finds a particularly advantageous, but not exclusive, application with PSA H2, O2 or CO2.

[0003] In general, a gas-phase adsorption process allows the separation of one or more molecules from a gaseous mixture containing them, by exploiting the difference in affinity of one or more adsorbents for the different constituent molecules of the mixture. The affinity of an adsorbent for a molecule depends on the structure and composition of the adsorbent, as well as on the properties of the molecule, particularly its size, electronic structure, and multipole moments. Examples of adsorbents include zeolite, activated carbon, activated alumina (possibly doped), silica gel, a carbon molecular sieve, a metal-organic structure, an oxide or hydroxide of alkali or alkaline earth metals, or a porous structure preferably containing a substance capable of reversibly reacting with the molecules, such as amines, physical solvents, metal complexing agents, or metal oxides or hydroxides.

[0004] The most common adsorbent materials are in the form of particles (beads, rods, crushed material, etc.) but also exist in structured forms such as monoliths, wheels, parallel passage contactors, fabrics, fibers, etc.

[0005] We can distinguish 3 main families of adsorption processes: lost charge processes, temperature modulation processes called TSA (Temperature Swing Adsorption) and finally PSA (Pressure Swing Adsorption) processes.

[0006] In lost charge processes, a new charge is put in place when the one in use is saturated by impurities or more generally when it can no longer play its protective role sufficiently.

[0007] In TSA type processes, the adsorbent at the end of use is regenerated in situ, that is to say that the impurities trapped are removed so that the said adsorbent recovers most of its adsorption capacity and can start a new purification cycle, the essential regeneration effect being due to a rise in temperature.

[0008] Finally, in PSA-type processes, the adsorbent at the end of the production phase is regenerated by desorption of impurities achieved through a decrease in their partial pressure. This pressure decrease can be obtained by lowering the total pressure and / or by purging with a gas that is free of or contains few impurities.

[0009] Pressure-modulated adsorption processes are used both to remove trace impurities, for example those present at less than one percent in the feed gas, and to separate mixtures containing tens of percent of different gases. In the first case, the process is generally referred to as purification (for example, gas drying), and in the second case, as separation (for example, the production of oxygen or nitrogen from atmospheric air).

[0010] FR2865554 Al describes a method for managing a pressure-modulated gas adsorption treatment unit operating with a high level of reliability and stability, particularly when one or more adsorbers are temporarily out of service due to maintenance or an unforeseen incident, such as a valve failure.

[0011] In the context of the present invention, the term PSA refers to any gas purification or separation process that employs a cyclic variation of the pressure experienced by the adsorbent between a high pressure, known as the adsorption pressure, and a low pressure, known as the regeneration pressure. Thus, this generic term PSA is used interchangeably to designate the following cyclic processes, which are also commonly given more specific names depending on the pressure levels involved or the time required for an adsorbent to return to its initial state (cycle time): VSA processes in which adsorption takes place substantially at atmospheric pressure, preferably between 0.95 and 1.25 bar abs and the desorption pressure is lower than atmospheric pressure, typically from 50 to 400 mbar abs; MPSA or VPSA processes in which adsorption takes place at a high pressure higher than atmospheric pressure, typically between 1.5 and 6 bar abs, and desorption at a low pressure below atmospheric pressure, generally between 200 and 600 mbar abs; PSA processes proper in which the high pressure is substantially greater than atmospheric pressure, typically between 3 and 50 bar abs and the low pressure substantially equal to or greater than atmospheric pressure, generally between 1 and 9 bar abs; RPSA (Rapid PSA) processes for which the duration of the pressure cycle is typically less than one minute; URPSA (Ultra Rapid PSA) processes for which the duration of the pressure cycle is on the order of a few seconds maximum.

[0012] It should be noted that these various designations are not standardized and that the limits are subject to variation.

[0013] An adsorber will therefore begin an adsorption period until it is saturated with the constituent(s) to be stopped at high pressure. It will then be regenerated by depressurization and extraction of the adsorbed compounds before being restored to a state ready to begin a new adsorption period. The adsorber has thus completed a pressure cycle, and the very principle of the PSA process is to chain these cycles one after the other; it is therefore a cyclic process. The time it takes an adsorber to return to its initial state is called the cycle time. In principle, each adsorber follows the same cycle with a time lag called the phase time, or simply the phase. We therefore have the following relationship: phase time = cycle time / number of adsorbers. We see that the number of phases is equal to the number of adsorbers.

[0014] This cycle typically includes periods such as: Production or Adsorption, during which the feed gas is introduced through one end of the adsorber, the most adsorbable compounds are preferentially adsorbed, and the gas enriched with the least adsorbable compounds (produced gas) is extracted through the other end. Adsorption can occur at increasing pressure, at a nearly constant pressure, or even at a slightly decreasing pressure. Depressurization, during which the adsorber, no longer supplied with feed gas, is evacuated through at least one end, removing some of the compounds contained in the adsorbent and the free volumes.Based on the direction of fluid flow during adsorption, we can define co-current, counter-current, or simultaneous co-current and counter-current depressurizations; Elution or Purge, during which a gas enriched in the least adsorbable constituents (purge gas) flows through the adsorbent bed to aid in the desorption of the most adsorbable compounds. Purge is generally performed counter-currently; Repressurization, during which the adsorber is at least partially repressurized before resuming an adsorption period. Repressurization can be counter-current and / or co-current; Dead time, during which the adsorber remains in the same state. These dead times can be an integral part of the cycle, allowing for the synchronization of steps between adsorbers, or they can be part of a step that has ended before the allotted time.The valves can be closed or remain in their current state, depending on the characteristics of the cycle.

[0015] Pressure swing adsorption (PSA) units are known to have a high number of adsorbents (N), commonly six or more, but sometimes exceeding twenty. Indeed, the current trend is to process increasingly higher flow rates, and it is generally much more advantageous to increase the number of adsorbents than to increase their individual volume. It is clear that the more adsorbents a unit has, the greater the risk that one of them will become inoperative.

[0016] The term "adsorber" should be understood here in its broadest sense, including all directly associated equipment such as valves and instrumentation. A failure in any of this equipment (the adsorber itself, valves, or instrumentation) will generally render the adsorber inoperable, meaning it can no longer be used in the cycle. This risk was taken into account as soon as PSA-type units were developed.

[0017] One initial solution was to implement a backup cycle for the nominal cycle involving N adsorbents. This involved programming a cycle from the outset that operated with N-1 adsorbents, with the possibility of isolating a failing adsorber from the unit—in practice, allowing the N adsorbents to be isolated independently. In fact, for systems with 6 or more adsorbers, the preferred solution was to divide the N adsorbents into isolable groups of n adsorbents and implement backup cycles operating with Nn, or possibly N-2n or N-3n, etc., adsorbents. This represents a compromise between performance, cycle complexity, and investment. Generally, n=2 was chosen, meaning the N adsorbents are grouped in pairs (N / 2 pairs). In this case, the number N of adsorbents must be even, and this solution is preferred for systems with 6, 8, 10, 12, 14, 16, etc. adsorbers.For units with a number of adsorbers that is a multiple of 3, for example 9, 12, 15, 18, 21, etc. adsorbers, the adsorbers can be grouped into groups of 3 adsorbers each and, for example, for a nominal cycle operating with 21 adsorbers, substitution cycles corresponding to 18, 15 and 12 adsorbers in operation can be provided.

[0018] We have mentioned here the division of N adsorbers into pairs or triplets, but it is conceivable to go further if, in the future, PSA units with a very large number of adsorbers were developed. This could be the case, for example, for URPSAs (Ultra Rapid PSAs) comprising a multitude of small adsorbers.

[0019] With this type of arrangement, a group (for example, a pair or triplet of adsorbers) is fluidically isolated from the unit when a failure is detected in at least one adsorber of the pair or triplet, such as a gas leak, a valve failure, or an instrumentation malfunction. Fluidic isolation means, for example, that the adsorber then exchanges no material flow with the rest of the unit, regardless of the stage of the pressure cycle. This isolation is generally achieved using isolation valves common to the group (for example, the pair or triplet).

[0020] When a group (for example, a pair or a triplet) of adsorbers is fluidically isolated, it is then possible to configure the unit (in reality, the control device) to operate with a reduced number of adsorbers (for example, N-2 or N-3, respectively). Substitution cycles with fewer adsorbers than the nominal cycle are pre-programmed into the control device, just like the nominal cycle.

[0021] It should be noted that for the most modern units, the detection of certain failures can be done automatically, as well as the switch from the nominal cycle to the substitution cycle with a reduced number of adsorbers.

[0022] A problem can arise when another adsorber, belonging to another group (for example, to another pair or triplet), becomes defective.

[0023] Isolating this other group requires setting up the unit to operate, for example, on the basis of N-4 (in the case where the groups are pairs of adsorbers) or N-6 (in the case where the groups are triplets of adsorbers) adsorbers, if such operation is still possible.

[0024] The output of this unit is then greatly reduced.

[0025] In the event that such operation (here, for example, with N-4 or N-6 adsorbers) is not possible or has not been planned, then the unit must be stopped.

[0026] These situations are not satisfactory, whereas in practice there are adsorbers of unity, namely in the two examples considered, one adsorber in each isolated pair, or two adsorbers in each isolated triplet, which are found to be unused even though they are in perfect working order.

[0027] The present invention aims to effectively overcome these drawbacks by proposing a method for managing a pressure-modulated adsorption gas treatment unit, the unit comprising: at least N adsorbers, N being a number greater than or equal to 6, the N adsorbers being arranged in G groups each comprising n=N / G adsorbers, n being an integer greater than or equal to 2, each adsorber comprising an instrumentation means, a control device allowing the unit to operate selectively according to a nominal cycle with N adsorbers and according to a degraded cycle, the degraded cycle excluding at least one group compared to the nominal cycle, a plurality of connections for access of the control device to the instrumentation means of each adsorber, characterized in that when a first group and a second group each comprise at least one operational adsorber and at least one faulty adsorber and when these first and second groups together comprise at least n operational adsorbers, said process then comprises the steps of: a) stopping the unit, b) fluidly isolating the faulty adsorbers from the unit,c) configure the connections so as to exchange the instrumentation of the faulty adsorber of the first group with the instrumentation of the operational adsorber of the second group, so that the first group again has, with respect to the control device, n operational adsorbers, d) configure the control device so that it controls the unit according to the degraded cycle, excluding the second group.

[0028] The invention thus makes it possible to optimize the gas production process of a pressure-modulated adsorption treatment unit when two adsorbers belonging to two distinct groups are defective. For example, when two adsorbers belonging to two distinct pairs are defective, or when up to three adsorbers belonging to two triplets are defective.

[0029] According to one embodiment, the instrumentation means is arranged to allow the measurement of information relating to a physical quantity and / or the control of an actuator of a valve.

[0030] According to one embodiment, the process includes step e) of putting into fluidic communication with the unit, each adsorber participating in the degraded cycle.

[0031] According to one embodiment, the unit includes at least one isolation valve to allow selective fluidic isolation and fluidic communication of at least one adsorber with respect to the unit or at least one group with respect to the unit.

[0032] According to one embodiment, the isolation valve includes an isolation valve for a group with respect to the unit

[0033] According to one embodiment, the isolation valve includes an isolation valve of an adsorber with respect to the unit.

[0034] According to one embodiment, each group is arranged to be selectively isolated, so that the isolation of one group allows the simultaneous fluidic isolation of the n adsorbers of the group from the rest of the unit.

[0035] According to one embodiment, the unit includes a plurality of cycle valves, to control the pressure cycle of each adsorber, each cycle valve being notably distinct from the isolation valve.

[0036] According to one embodiment, the instrumentation means includes at least one sensor and / or a valve.

[0037] According to one embodiment, the sensor includes a pressure sensor and / or differential pressure sensor and / or temperature sensor.

[0038] According to one implementation, the instrumentation means includes a control of a valve.

[0039] According to one design, the unit contains between 6 and 24 adsorbers.

[0040] According to one design, the unit is arranged to produce either dihydrogen or carbon dioxide.

[0041] The invention further relates to a gas treatment unit by pressure-modulated adsorption, the unit comprising: at least N adsorbers, N being a number greater than or equal to 6, the N adsorbers being arranged in G groups each comprising n=N / G adsorbers, n being an integer greater than or equal to 2, each adsorber comprising an instrumentation means, a control device enabling the unit to operate selectively according to a nominal cycle with N adsorbers and according to a degraded cycle, the degraded cycle excluding at least one group in relation to the nominal cycle, a plurality of links for access of the control device to the instrumentation means of each adsorber, the control device comprising a microprocessor arranged for the implementation of steps a) to d) of the process as described above.

[0042] In other words, the microprocessor is arranged for the implementation of each of the steps a) to d) of the process as described above.

[0043] According to one embodiment, the microprocessor is further arranged for the implementation of step e) of the process as described above.

[0044] According to one implementation, the links are configured to be interchangeable, in particular between a failing adsorber of one group and an operational adsorber of another group.

[0045] All the features described above are applicable alone or in combination with this latest invention.

[0046] The invention further relates to a computer program product comprising instructions which lead the unit as described above to perform steps a) to d) of the process as described above, when the program is executed by the microprocessor.

[0047] According to one embodiment, the computer program product includes instructions that lead the unit as described above to execute step e) of the process as described above, when the program is executed by the microprocessor.

[0048] The invention also relates to a gas treatment unit by pressure-modulated adsorption, the unit comprising: at least N adsorbers, N being a number greater than or equal to 6, the N adsorbers being arranged in G groups each comprising n=N / G adsorbers, n being an integer greater than or equal to 2, each adsorber comprising an instrumentation means, a control device allowing the unit to operate selectively according to a nominal cycle with N adsorbers and according to a degraded cycle, the degraded cycle excluding at least one group from the nominal cycle, a plurality of links for access of the control device to the instrumentation means of each adsorber, the unit being configured by the control device to be in the degraded cycle, the unit comprising a first group and a second group each comprising at least one operational adsorber and at least one failed adsorber, all adsorbers participating in the degraded cycle being in fluidic communication with the unit and all failed adsorbers being fluidically isolated from the unit.

[0049] According to one embodiment, the links between a failing adsorber of one group and an operational adsorber of another group are exchanged.

[0050] The invention will be better understood upon reading the following description and examining the accompanying figures. These figures are given only to illustrate, but in no way limit, the invention. In particular, they only describe the case where the adsorbers of the unit are arranged in groups of two (i.e., in pairs), but the same principle could apply if the adsorbers were grouped in triplets, quadruplets, etc. [ Fig.1 ] There [ Fig.1 ] is a schematic representation of a PSA unit in which the process according to the invention is implemented; [ Fig.2 ] There [ Fig.2 ] is a schematic representation of the PSA unit of the [ Fig.1 ], in which a group is isolated from the unit via its isolation valve; [ Fig.3 ] There [ Fig.3 ] is a schematic representation of the PSA unit of the [ Fig.2 ], in which an adsorber from another group is non-operational; [ Fig.4 ] There [ Fig.4 ] is a schematic representation of the PSA unit of the [ Fig.3 ], implementing the method according to the invention. Fig.5 ] There [ Fig.5 ] is a schematic representation of the permutation of the links between an adsorber and the control device to replace a non-operational adsorber with an operational adsorber in a pair.

[0051] It is necessary, firstly, to clarify the terms used to describe the state of an adsorber as well as what can be understood by interchanging (or permuting) the means of instrumentation between two adsorbers.

[0052] An adsorber (in the broadest sense, including its equipment, notably valves and directly connected instrumentation) is considered operational if its condition allows it to be integrated into the unit's operating cycle. An operational adsorber generally participates in the unit's operation, but it can also be unused or inactive, in which case it is isolated from the unit. It remains available, however, to be integrated into the unit's cycle if needed.

[0053] Conversely, an adsorber will be non-operational if its condition does not allow it to be integrated into the cycle. It is then referred to as a failing or defective adsorber.

[0054] An adsorber, whether faulty or simply unused, will generally be isolated from the rest of the unit. This isolation can be achieved by: by means of an isolation valve relative to the group to which it belongs (in this case, the other adsorbers belonging to the same group are also isolated from the unit); or by means of closing all of its cycle valves; or by the implementation of specific means of isolation such as solid flanges, plugs, a bezel seal, a manual valve (these specific means are generally implemented for a long shutdown or during an intervention by an operator on the adsorber).

[0055] It should be noted that when an adsorber is isolated individually, i.e. by its own means (cycle valves, specific means), the other adsorbers in its group, if they are operational, can be put into fluidic communication with the unit and participate in the cycle if this is useful, as soon as the isolation valve relating to the group in question is open.

[0056] In some cases, each adsorber in a group may have its own isolation valve, which will generally replace the group's isolation valve. Thus, the isolation valves of each adsorber can be controlled simultaneously and, in this case, perform the same function as a single isolation valve for a group of adsorbers.

[0057] An isolation valve is a valve that isolates a group of adsorbers, or a valve that isolates a single adsorber. Such an isolation valve is distinct from a cycle valve.

[0058] Regarding the reversal of instrumentation methods, the terms used are general given the diversity of possible connections between local instrumentation linked to and in the immediate vicinity of an adsorber and the control device located in the control room, or at least remotely, which processes the information. A non-exhaustive example of such a connection is described below.

[0059] We know that a sensor measures a physical quantity (pressure, flow rate, temperature, etc.) and that its transmitter translates this into a usable quantity, often an analog signal that can travel through electrical wires (also called "instrument cables"). These wires connect to the terminal blocks of local junction boxes, located near the site equipment, which group the connections from several sensors. From these local junction boxes, a multi-wire instrumentation cable runs, grouping all the inputs, for example, those from a dozen sensors. These instrumentation cables, the number of which depends on the complexity of the unit, run via a cable tray to the control room's technical room, and more specifically to the PSA cabinet. At this point, the wires for the different sensors are separated and connected to the cabinet's terminal blocks.These terminal blocks are also connected to the input / output cards which transform the signals that were assumed to be analog until then into digital signals used by the control device.

[0060] Regarding cycle valves, it is generally the valve positioner that receives information from the control device and drives the actuator, which in turn moves the valve disc via a mechanical system (stem, gear, etc.). In return, it can transmit the valve position, alarms, limit switch information, etc., back to the control device.

[0061] "Instrumentation means", for example, refers to the information transmission system between the site and the control room, whether it involves sensors or control valves.

[0062] Local junction boxes can group wires from sensors of the same type, for example, all pressure sensors or those related to valves. All information related to an adsorber can also be grouped together. These junction boxes can be separate or grouped in one or more panels.

[0063] It is understood that the reversal of an instrumentation means between two adsorbers can be done at different points in the information path circuit: local terminal block, technical cabinet terminal block, input / output cards, connection with the control device.

[0064] The final choice depends on the configuration of each system. Note that grouping information by adsorber can facilitate this reversal.

[0065] The most common type of analog connection has been briefly described above. Signals can also be transmitted digitally or via wireless networks. It is always possible to swap the signals from two adsorbers by adapting the procedure to the technology used.

[0066] In all cases, the result will be that the control device, via its program, will receive information from a given adsorber (in reality not operational) and will transmit orders and instructions to this adsorber, while in practice it will control another adsorber of the unit.

[0067] With reference to the [ Fig.1 [ ] A PSA 100 unit is shown, implementing N = 12 adsorbers 1-12. This PSA 100 unit is configured to include 6 groups (here, pairs) 21-26 of adsorbers (pair 21 containing adsorbers 1 and 2, pair 22 containing adsorbers 3 and 4, etc.). Each pair 21-26 of the PSA 100 unit is arranged to be selectively isolated, such that isolating one pair 21-26 allows for the fluidic isolation of the two adsorbers in the pair, from the unit 100. More precisely, each pair 21-26 can be isolated from the rest of the unit 100 by an isolation valve 31-36. In the example shown, each isolation valve 31-36 is an isolation valve of the group from the unit, that is to say, it allows the group of adsorbers to be isolated from the unit.

[0068] The choice to group the adsorbers 1-12 into groups (here, in pairs) 21-26 for the isolation of at least one adsorber 1-12 of the unit is, as explained, a compromise between the flexibility of use and the complexity of the technical installations of the PSA 100 unit.

[0069] Unit 100 further comprises a plurality of cycle valves and a control device 200 for operating the cycle valves to ensure pressure cycling. Each adsorber 1-12 is thus controlled by cycle valves attached to it.

[0070] Unit 100 also includes a plurality of links 45, 47, 48 for access to instrumentation means for each adsorber 1-12.

[0071] More specifically, these links 45, 47, 48 include electrical cables, connecting the instrumentation means of each adsorber 1-12 to the control device 200.

[0072] Instrumentation equipment includes, for example, pressure sensors and cycle valve switches. Switches here refer to any means specific to a valve that allows its control and operation during each stage of the cycle.

[0073] Thus, the control device 200 can read the information delivered by the sensors and control the cycle valves of each adsorber 1-12.

[0074] There [ Fig.2 ] represents the same unit as that of the [ Fig.1 ], but operating according to a cycle implementing only 10 adsorbers, pair 23 having been removed from the cycle following a malfunction of adsorber 6.

[0075] This type of operation was planned from the unit's design stage, and the switch from 12 to 10 adsorbers was automatic. Adsorbers 5 and 6 are isolated from the rest of unit 100 because the group consisting of these two adsorbers was isolated via a group isolation valve. Alternatively, isolation can be achieved via the individual isolation valves of each adsorber.

[0076] The isolation valves 31-36, which allow each of the groups (here, each of the pairs) 21-26 to be isolated respectively, are symbolically represented. On the [ Fig.2 ], the isolation valve 33 is closed and isolates the adsorbers 5 and 6 from the unit.

[0077] While operating in this configuration, a second adsorber (namely adsorber 8 of pair 24, in the example shown in the [ Fig.3 ]) presents a failure. This situation is represented on the [ Fig.3 ]. If the unit was not designed to operate with two fewer groups (here, two fewer pairs), either the PSA unit detects the fault and stops automatically, or the PSA unit must be stopped manually.

[0078] If, on the other hand, the unit was designed to operate with only 8 adsorbers, it could operate according to this last cycle but with very significantly degraded performance.

[0079] None of these solutions is satisfactory, especially if restoring at least one of the two non-operational adsorbers requires significant intervention time.

[0080] The method according to the invention will then consist of interchanging the instrumentation means of the faulty adsorber 8 with those of the operational adsorber 5 at the control device 200 in order to form a new operational pair 24. The control device 200 will thus, in practice, control the adsorber 5 instead of the adsorber 8 without having to intervene in the programming of the cycles or physically move any piping. In other words, the control device 200, believing it is controlling the adsorber 8, will in fact control the adsorber 5.

[0081] There [ Fig.4 ] symbolizes this new configuration in which pair 24 still exists for control device 200 but now includes adsorber 5 instead of adsorber 8.

[0082] There [ Fig.5 ] schematically represents the control device 200, the links 45, 47, 48 between respectively the adsorbers 5, 7, 8 and the corresponding inputs / outputs 55, 57, 58 located on one of the faces 201 of the control device 200.

[0083] Inputs / outputs 55, 56 correspond to pair 23, while inputs / outputs 57, 58 correspond to pair 24. By disconnecting link 48 and replacing it with link 45, the control device 200 will control the pressure cycle of the adsorber 5 which has been swapped with the adsorber 8, in the same way that it would control the adsorber 8.

[0084] It may be possible to connect link 48 to input / output 55 but pair 23 being excluded from the cycle, there will be no corresponding action.

[0085] We have assumed here that the connections are made via electrical cables. They could just as easily be optical fibers or radio links such as Wi-Fi, Bluetooth, or equivalent. Regardless of the type of connection, the goal is to form a new group (pair 24 in the example of the [ Fig.5 ]) which can be controlled by control device 200 without requiring modification of the programming of control device 200 or the piping. Instead, the inputs / outputs (48, in the example of the [ Fig.5 ]) normally dedicated to a failing adsorber (8, in the example of the [ Fig.5 ]) for another functional adsorber of unit 100 (5, in the example of the [ Fig.5 ]).

[0086] This isolation is carried out in compliance with the safety rules specific to the sites, depending on the pressures involved and the constituents present in the adsorbers.

[0087] You can use bezel gaskets, solid flanges or plugs.

[0088] Isolated adsorbers can be depressurized, purged or inertized as appropriate.

[0089] The restart itself may involve pressurization steps prior to switching to automatic mode. It may then be necessary to force open the isolation valves 33 and 34 of pairs 23 and 24 respectively beforehand.

[0090] The method recommended here requires modifications to the instrumentation connections which may involve, for example, depending on the PSA unit, two to five valves, or one or two pressure or differential pressure sensors, or even a temperature sensor.

[0091] Finally, it should be noted that the arrangement of the adsorbers by group is both physical and logical. For example, two adsorbers in a particular group will have their inlet / outlet pipes connected to specific manifolds, and these manifolds will have their own isolation valves. The control system, for its part, is capable of controlling two adsorbers within a given group and distinguishing one group from other adsorber groups.

Claims

1. A method for managing a failure of a gas treatment unit (100) by pressure swing adsorption, the unit (100) comprising: at least N adsorbers (1-12), N being a number greater than or equal to 6, the N adsorbers being arranged in G groups (21-26) each comprising n=N / G adsorbers, n being an integer greater than or equal to 2, each adsorber (1-12) comprising an instrumentation means, a control device (200) allowing the unit (100) to operate selectively according to a nominal cycle with N adsorbers and according to a degraded cycle, the degraded cycle excluding at least one group (21-26) with respect to the nominal cycle, a plurality of links (45, 47, 48) for access by the control device (200) to the instrumentation means of each adsorber (1-12), characterized in that when a first group (G1) and a second group (G2) each comprise at least one operational adsorber (1-12) and at least one faulty adsorber (6,8) and when these first (G1) and second (G2) groups together comprise at least n operational adsorbers, said method then comprises the steps of: a) stopping the unit (100), b) fluidically isolating the faulty adsorbers (1-12) from the unit (100), c) configuring the links (45, 47, 48) so as to swap the instrumentation means of the faulty adsorber of the first group (G1) with the instrumentation means of the operational adsorber of the second group (G2), so that the first group (G1) again comprises, with respect to the control device (200), n operational adsorbers, d) parameterizing the control device (200) so that it controls the unit (100) according to the degraded cycle, by excluding the second group (G2).

2. The method according to the preceding claim, comprising the step of: e) placing in fluid communication with the unit (100) each adsorber (1-12) participating in the degraded cycle.

3. The method according to one of the preceding claims, the unit (100) comprising at least one isolation valve (31-36) to selectively allow fluidic isolation and placement in fluid communication of at least one adsorber with respect to the unit (100) or at least one group with respect to the unit (100).

4. The method according to the preceding claim, the unit (100) comprising a plurality of cycle valves to control the pressure cycle of each adsorber (1-12), each cycle valve being notably distinct from the isolation valve (31-36).

5. A gas treatment unit (100) by pressure swing adsorption, the unit (100) comprising: at least N adsorbers (1-12), N being a number greater than or equal to 6, the N adsorbers being arranged in G groups (21-26) each comprising n=N / G adsorbers, n being an integer greater than or equal to 2, each adsorber (1-12) comprising an instrumentation means, a control device (200) allowing the unit (100) to operate selectively according to a nominal cycle with N adsorbers and according to a degraded cycle, the degraded cycle excluding at least one group (21-26) with respect to the nominal cycle, a plurality of links (45, 47, 48) for access by the control device (200) to the instrumentation means of each adsorber (1-12), the control device (200) comprising a microprocessor configured to implement the steps of the method according to one of the preceding claims.

6. The unit (100) according to claim 5, characterized in that the links (45, 47, 48) are configured to be interchangeable, notably between a faulty adsorber of one group and an operational adsorber of another group.

7. A computer program product comprising instructions which cause the unit (100) according to claim 5 or 6 to execute the steps of the method according to one of claims 1 to 4, when the program is executed by the microprocessor.