Method for controlling a fuel cell, fuel cell and associated vehicle

Abstract

Fuel cell control method, fuel cell and associated vehicle. This fuel cell control method is implemented during the fuel cell start-up phase. The fuel cell comprises a stack of several unit cells distributed into several subgroups, and a control device configured to measure a unit voltage across the terminals of each subgroup.While the unit voltage of each subgroup is initially zero and then increases to an operating voltage (UF), the method includes measuring the associated unit voltage (Uc) for each subgroup, identifying one or more so-called "fast" subgroups whose unit voltage first reaches a predetermined, non-zero voltage threshold (U1) lower than the operating voltage, and then a regulation step in which at least one of the fast subgroups is placed in a coupled configuration, where the subgroup in question is connected to a respective electrical load, so as to limit the increase in the unit voltage of the subgroup in question. (See Figure 4 for the abbreviation.)

Description

Title of the invention: Method for controlling a fuel cell, fuel cell and associated vehicle

[0001] The present invention relates to a method of controlling a fuel cell, in particular during its start-up, a fuel cell configured to implement such a control method and a vehicle comprising such a fuel cell.

[0002] A fuel cell is a device that generates electricity through an electrochemical reaction between a fuel, for example dihydrogen, and an oxidant, for example oxygen from the air. This discussion focuses on proton exchange membrane fuel cells with a solid electrolyte – also known as PEMFCs – which typically comprise a stack of several unit cells, each constituting an electrochemical generator.

[0003] Schematically, each unit cell comprises two separators, also called polar plates, between which a solid electrolyte in the form of a proton exchange membrane is interposed. The membrane is made, for example, of a sulfonated perfluorinated polymer material. Within each cell, each separator, together with the corresponding membrane, delimits a reactive compartment. The membrane is thus interposed between two reactive compartments. One of the two reactive compartments, called the cathodic compartment, houses a cathodic element, formed by a cathodic catalytic layer located on the surface of the membrane, while the other reactive compartment, called the anodic compartment, houses an anodic element, formed by an anodic catalytic layer located on the surface of the membrane. The catalytic layer of the membranes generally comprises carbon, for example in the form of a carbon fiber felt.

[0004] The assembly of the membrane and the anodic and cathodic catalytic layers forms a membrane-electrode assembly, generally known by the acronym AME in French, or MEA in English.

[0005] For two adjacent cells, a separator from one of the two cells is placed back-to-back with a separator from the other cell. These two separators together form a bipolar separator, also called a bipolar plate. A cooling compartment, through which a cooling fluid such as glycol water circulates, is generally arranged between the two separators of the bipolar separator. Each reactive compartment also generally includes a gas diffusion layer, located between the bipolar separator and the catalytic layer. allowing good circulation of fuel or oxidizer from the separator to the catalytic layer.

[0006] Dihydrogen, air, and—where applicable—coolant are referred to as "operating fluids," which are supplied to the fuel cell during its operation. Dihydrogen and air are reactants, while the coolant does not participate in the electrochemical reaction. Depending on the operating phases of the fuel cell, the supply of one or more of the operating fluids is continuous, intermittent, or even interrupted.

[0007] The fuel cell provides openings to supply fluids to each of the reactive compartments and to the fluids between two adjacent cells. In a widely used design, each bipolar separator supplies, on one side, the fuel to the cell adjacent to that side and, on the other side, the oxidant to the cell adjacent to that other side, the supplies provided by the bipolar separators being in parallel.

[0008] Generally, in a unit cell, the cathode compartment is supplied with an oxidant, for example oxygen, most often in the form of an oxygen-containing air supply, and the anode compartment is supplied with a fuel, for example dihydrogen. When the stack is formed, the cathode compartments of the unit cells communicate with each other, step by step, forming a cathode circuit of the stack. Similarly, the anode compartments of the unit cells communicate with each other, step by step, forming an anode circuit of the stack.During the operation of the fuel cell, the anodic circuit of the stack is supplied with hydrogen by an anodic supply circuit of the fuel cell, while the cathodic circuit of the stack is supplied with air - and therefore oxygen - by a cathodic supply circuit of the fuel cell, and the stack is supplied with coolant by a cooling circuit.

[0009] When the fuel cell is operating, at the anode, the dihydrogen used as fuel is ionized to produce protons that pass through the membrane. The electrons produced by this reaction pass through an electrical circuit external to the cell to form an electric current. At the cathode, oxygen is reduced and reacts with the protons to form water. This electrochemical reaction thus creates an electrical potential difference between the two separators of each unit cell. The fuel cell therefore includes an electrical isolation device, designed to prevent any electrical contact between two adjacent bipolar separators and between each unit cell and the external environment. as well as a sealing device to prevent leaks of operating fluids, in particular to prevent the fluid circulating in a reactive compartment from contaminating a neighboring reactive compartment.

[0010] The electrical potential difference between the two separators of each unit cell creates a voltage across the terminals of each unit cell, referred to as the "unit voltage". All the cells of the fuel cell are electrically connected together in series, so that the voltage delivered across the stack is equal to the sum of the unit voltages of all the unit cells.

[0011] In particular, when the fuel cell is completely shut down and depolarized, the anodic feed circuit is closed, and gases other than hydrogen, notably air containing oxygen, nitrogen, etc., tend to enter the anodic circuit of the stack. When the fuel cell is restarted, the anodic feed circuit is reopened, along with the opening of the anodic purge solenoid valve, so as to purge the anodic circuit until it contains essentially only hydrogen. This purging step is also called a flush. Initially, the cathodic feed circuit is kept closed during this purging step, preventing the fuel cell from repolarizing. The cell voltage is thus considered to remain zero, or substantially zero.

[0012] Once the anodic circuit has been purged, the cathodic circuit is in turn supplied with air, the electrochemical reaction gradually takes place on both sides of the membrane of each unit cell, and the unit voltage of each unit cell gradually increases, until it reaches a substantially stable value equal to an operating voltage, which is typically on the order of 1 V - Volt -.

[0013] However, the evolution of the unit voltage of each unit cell of the stack is not perfectly synchronized, so that a voltage differential generally exists between the unit cells of the stack during start-up, which subjects the unit cells to a risk of corrosion, in particular corrosion of the carbon present on the catalytic layer of the membranes, directly impacting the life of the fuel cell.

[0014] US2008 / 032163 (Al) describes, for example, a method of startup protection of the fuel cell, in which the entire stack is connected to an electrical load, in order to limit the increase in voltage across the stack. The electrical potential difference between the individual cells is thus limited, reducing the risk of corrosion. A disadvantage of this approach is that the power to be dissipated is significant, potentially reaching several kilowatts, which leads to substantial heating of the fuel cell. Such heating is not not desired, particularly with regard to safety issues and battery life.

[0015] The invention specifically aims to address these problems by proposing a method for controlling a fuel cell that reduces the risk of corrosion during fuel cell start-up while also reducing fuel cell heating. The invention thus increases the fuel cell's lifespan (or durability).

[0016] To this end, the invention relates to a method for controlling a fuel cell, the fuel cell comprising: - a stacking of several unit cells, the unit cells being distributed into several juxtaposed subgroups, each subgroup comprising at least one unit cell, and - a control device, which is configured to measure a unit voltage, which is equal to an electrical voltage across the terminals of each subgroup.

[0017] During a start-up phase of the fuel cell, when the unit voltage of each subgroup is initially zero and then increases until it reaches a voltage substantially equal to an operating voltage, the method comprises:

[0018] a) for each subgroup, the measurement of the associated unit voltage by means of the control device, so as to monitor the evolution of the unit voltage as a function of time,

[0019] b) an identification step, during which one or more of the subgroups, called fast subgroups, are identified, whose unit voltage first reaches a first predetermined voltage threshold, the first voltage threshold being non-zero and lower than the operating voltage,

[0020] c) a regulation step, during which at least one of the fast subgroups is put into a coupled configuration, in which the subgroup in question is connected to a respective electrical load, so as to limit the increase in the unit voltage of the subgroup in question.

[0021] Thanks to the invention, the fast subgroup whose unit voltage reaches the first voltage threshold is connected to the corresponding electrical load, which consumes electrical energy and reduces the rate of increase of the unit voltage. A voltage differential between said fast subgroup thus connected to the corresponding electrical load and the other subgroups thus tends to decrease, which reduces the risk of corrosion within the stack and helps to extend the service life of the fuel cell. The number of subgroups thus connected to the corresponding electrical charge is limited, which also limits the amount of energy dissipated, and therefore limits the heating of the fuel cell.

[0022] According to advantageous but not mandatory aspects of the invention, such a piloting method may incorporate one or more of the following features taken individually or in any technically permissible combination: - During the regulation stage, as long as a proportion of the number of fast subgroups relative to a total number of subgroups in the stack is less than a first predetermined percentage, then each fast subgroup is connected to the associated electrical load. - The first percentage is between 20% and 80%, preferably less than 70%, preferably even less than 60%. - Once the proportion of fast subgroups connected to the corresponding electrical load is equal to the first percentage, no further subgroups are connected to the corresponding electrical load. - Each fast subgroup, once connected to the associated electrical load, is disconnected from the associated electrical load no later than after a first predetermined time interval. - We define an average voltage rise delay, which is equal to the time elapsed between the moment when the unit voltage of the first fast subgroup reaches a second voltage threshold and the moment when the unit voltage of half of the subgroups reaches the second voltage threshold, the second voltage threshold being less than or equal to the first voltage threshold, while the first time interval is calculated as a function of the average voltage rise delay. - The second voltage threshold is equal to the first voltage threshold. - The first time interval is equal to a coefficient multiplied by the average delay, the coefficient being less than 2, preferably greater than 1, preferably still between 1.2 and 1.8. - The first voltage threshold is between 0.05 V and 0.6 V, preferably between 0.05 V and 0.4 V, preferably also equal to 0.1 V ± 0.05 V.

[0023] According to another aspect, the invention relates to a fuel cell, which comprises: - a stack of several unit cells, the unit cells being distributed into several juxtaposed subgroups, each subgroup comprising at least one unit cell, - a control device, which is configured to measure a unit voltage of each subgroup, and which comprises several electrical loads, each electrical load being associated with a respective subgroup and being selectively connectable, reversibly, to the associated subgroup,

[0024] in which the fuel cell is configured to implement the piloting method as defined above.

[0025] Advantageously: - The control device also includes a switching device between each subgroup and each associated electrical load. - Every switching device is a switch. - Every electrical charge is a resistance. - The control module is configured to perform an impedance measurement across the terminals of the corresponding subgroup using a four-wire method and the corresponding resistance.

[0026] The invention finally relates to a vehicle, comprising a fuel cell as described above.

[0027] The invention will be better understood, and other advantages thereof will become more apparent from the following description of an embodiment of a fuel cell control method, a fuel cell and an associated vehicle, in accordance with its principle, given solely by way of example and with reference to the accompanying drawings, in which:

[0028] - [Fig. 1] [Fig. 1] schematically represents, on an inset a), a vehicle, in accordance with a first embodiment of the invention, the vehicle comprising a fuel cell, also in accordance with the invention, on an insert b), a housing of the fuel cell of the insert a), and on an insert c), a stack of unit cells of the fuel cell of the insert a), the stack being received in the housing of the insert b);

[0029] - [Fig.2] [Fig.2] represents respectively, on two inserts a) and b), the stacking of the [Fig. 1] surmounted by a control device, and a schematic representation of the stack and the control device,

[0030] - [Fig.3] [Fig.3] schematically represents, on two inserts a) and b), the evolution voltage measurements across the terminals of the unit cells of the stack during a start-up phase of the fuel cell, when the control method of the invention is not implemented.

[0031] - [Fig.4] [Fig.4] represents, on two inserts a) and b), the principle of the process of piloting the invention,

[0032] - [Fig.5] [Fig.5] represents the evolution of the voltage measurements across the terminals of unit cells during a start-up phase of the fuel cell of [Fig.1] when the control method according to the invention is implemented.

[0033] - [Fig.6] [Fig.6] is a flowchart schematically illustrating a mode of implementation of the method for controlling the invention.

[0034] A vehicle 10 is shown in [Fig. 1]. Vehicle 10 is here a road vehicle, such as a car, truck, or forklift. Alternatively, vehicle 10 is a railway vehicle such as a train or subway, or even an aircraft or ship.

[0035] The vehicle 10 here comprises wheels 12, by means of which the vehicle 10 rests on a ground 14. The ground 14 is assumed to be substantially horizontal, the following description being given in relation to the orientation of the various elements of the vehicle 10 as shown in the figures, bearing in mind that this may be otherwise in reality. The vehicle 10 comprises a fuel cell 20. The fuel cell 20, also simply referred to as cell 20 in this description, moves with the vehicle 10 and is a so-called "on-board" fuel cell, that is to say, mounted on a chassis of the vehicle 10, as opposed to so-called "stationary" fuel cells which are fixed, placed on the ground or in a building.

[0036] The fuel cell 20 is fueled by a fuel, in this case hydrogen, which reacts with oxygen from the air to generate electricity for the operation of the vehicle 10, with some of the fuel's energy being dissipated as heat. Specifically, the vehicle 10 includes, for its propulsion, an electric motor, which is powered, directly or indirectly (via a battery, for example), by the electrical energy generated by the fuel cell 20. The electric motor is not shown.

[0037] With reference to [Fig. 1]#a), the fuel cell 20 comprises a casing 22, also called a housing, the casing 22 comprising a sleeve 24. The sleeve 24 has a hollow shape, for example that of a hollow parallelepiped, extending along a principal axis A20 and which here has a cylindrical shape with a substantially rectangular cross-section. The sleeve 24 provides an internal compartment V24 with two opposing end openings, the end openings being sealed respectively by two covers 26A and 26B. The two covers 26A and 26B and the sleeve 24 thus define the internal compartment V24, which is also, by extension, an internal compartment of the casing 22. The covers 26A and 26B each extend along planes orthogonal to the principal axis A20. The sleeve 24 and the covers 26A and 26B are preferably made of electrically insulating material, such as a polymer material, possibly fiber-reinforced.Alternatively, covers 26A and 26B are coated, at least on one inner face, with an electrically insulating material.

[0038] Fluid passages 28 are provided through the casing 22 to allow the passage of fluids for the operation of the fuel cell 20. The operating fluids here comprise three fluids, including two gaseous fluids, in this case air - containing dioxygen - and dihydrogen, and a dielectric heat transfer fluid, for example The liquid, in this case glycol water, is used. The fluid passages are formed by fluid connections, which are located here on cover 26A, situated on the right of [Fig. 1]#b). Alternatively, all or part of the fluid connections are located on cover 26B. Alternatively, all or part of the fluid connections are located on sleeve 24.

[0039] In the illustrated example, the stack 20 comprises three pairs of fluid connections 28, each pair being intended for the circulation of a specific operating fluid. The three pairs of fluid connections 28 include a first pair of connections 28A, a second pair of connections 28B, and a third pair of connections 28C. For each pair of connections, one of the connections, referred to as the "inlet connection," is intended for the intake of the corresponding operating fluid, while the other connection, referred to as the "outlet connection," is intended for the extraction of the corresponding operating fluid. In the figures, the direction of flow of the operating fluids is schematically represented by arbitrarily oriented arrows, although this may be different in reality. The size and arrangement of the fluid connections 28 are not limiting.

[0040] The stack 30 of the fuel cell 20 is schematically represented in partially exploded perspective in inset c) of [Fig.1].

[0041] The stack 30 is held between two end plates 31A and 31B, which are schematically represented in [Fig.1]#a) and in Figures 2 and 3. The end plates 31A and 31B make it possible in particular to hold the stack 30 in the housing 22, and to supply the stack 30 with fuel, in the example of dihydrogen in gaseous form, and with oxidizer, in the example of air in gaseous form, and, where applicable, the circulation of a heat transfer fluid for a cooling circuit.

[0042] The invention will be described more particularly in the context of a common construction in which the stack 30 is formed of several unit cells 32, each comprising a membrane-electrode assembly 34 and two bipolar plates 36, arranged on either side of the membrane-electrode assembly 34. The membrane-electrode assembly 36 is also referred to as MEA 36 in the context of this description. However, the invention is also applicable in the context of fuel cells of another type, in particular of the solid electrolyte ion-exchange membrane type. The membrane-electrode assemblies 34 and the bipolar plates 36 are each substantially planar and are stacked along a stacking axis A30, which is considered to coincide with the main axis A20 when the stack 30 is received in the casing 24.

[0043] When the fuel cell 20 is installed in the vehicle 10, in a normal operating configuration, the fuel cell 20 is stationary relative to the vehicle 10. Thus, the main axis A20 is considered to be substantially horizontal, while the housing 22 has a top side 21A and a bottom side 21B, which is opposite the top side 21A and which is located between the top side 21A and the ground 14. In the illustrated example, the covers 26A and 26B are considered to extend along substantially vertical planes.

[0044] It is assumed that, for a given fuel cell 20, all the unit cells 32, also simply called cells 32, are identical to each other, and therefore have identical characteristics. In practice, each bipolar plate 36 is positioned between two consecutive cells 32 and is common to both of these consecutive cells 32. Each bipolar plate 36 comprises a first face 16A, called the anode face, which supplies one of the two cells 32 with dihydrogen, and a second face 16B, which is opposite the first face 16A and is called the cathode face, which supplies the other of the two cells 32 with air. In other words, a cell 32 is supplied with dihydrogen by a first bipolar plate 36 and with air by a second bipolar plate. Since air contains oxygen, the cell 32 is thus supplied with oxygen.

[0045] In the following description, the terms oxygen and dioxygen, as well as hydrogen and dihydrogen, are used interchangeably.

[0046] In the example, each bipolar plate 36 is formed by the assembly of two half-plates, which form channels for the circulation of the heat transfer fluid. The circulation of this heat transfer fluid plays no role in the electrochemical reactions of the fuel cell 10, but allows the temperature of the cells 32 to be controlled. In normal operation of the fuel cell 20, the fuel cell 20 generates electricity while releasing heat, which is dissipated by circulating the heat transfer fluid. The heat transfer fluid then acts as a coolant.

[0047] Each cell 32 is divided into an anodic compartment, formed between the bipolar plate 36 supplying the cell with dihydrogen and the membrane 34, and a cathodic compartment, formed between the bipolar plate 36 supplying the cell with air and the membrane. Generally, in each unit cell 32, the cathodic compartment is supplied with an oxidant, for example oxygen, here in the form of an oxygen-containing air supply, and the anodic compartment is supplied with a fuel, for example dihydrogen. When the stack 30 is formed, the cathodic compartments of the unit cells 32 communicate with each other, step by step, forming a cathodic circuit 38A of the stack. Similarly, the anodic compartments of the unit cells communicate with each other, step by step, forming an anodic circuit 38B of the stack.Similarly, the cooling compartments of the unit cells 32 communicate with each other. with the others, step by step, forming a 38C cooling circuit of the stack 30.

[0048] With reference to [Fig. 2], the fuel cell 20 also includes a control device 40, which is mounted here on one side of the stack 30 and connected to the bipolar plates 36. In the non-limiting example shown, the control device 40 is advantageously modular, i.e., it is composed of several identical modules 42, which are juxtaposed and connected to the bipolar plates 36. In [Fig. 2]#a), one of the two modules 42 is partially shown to reveal the connection to the bipolar plates 36. The modules 42 are electrically connected to the bipolar plates 36 via pins 44, which are received in pockets 46 formed on the edge of the bipolar plates 36. Other methods of connection between the control device and the bipolar plates 36 are of course possible.

[0049] The control device 40 is configured to acquire an electrical potential from the bipolar plates 36 to which the control device 40 is connected. For two given bipolar plates 36, an electrical voltage between these two bipolar plates 36 is calculated by taking the difference between the electrical potentials associated with each of these bipolar plates 36. The electrical voltage between two bipolar plates 36 is advantageously measured using an impedance measurement method known as the "four-wire" or "Kelvin method." An electrical current of known value is injected across the terminals of the cell or group of cells considered, generating an electrical voltage response that is measured, for example, using a voltmeter. For each unit cell 32, the potential difference between the two bipolar plates 36 associated with this unit cell 32 is called the "unit voltage" Uc, expressed in Volts.The control device 40 is thus configured to measure the unit voltage Uc across the terminals of each unit cell 32.

[0050] With reference to [Fig. 2]#b), the stack 30 and the control device 40 are shown schematically and partially in cross-section. The control device 40 comprises several electrical loads 50, each electrical load 50 being associated with a respective unit cell 32 and being selectively connectable or disconnectable from the associated unit cell 32.

[0051] In the illustrated example, the electrical charges 50 are advantageously electrical resistors, the value of which is selected according to the power to be dissipated, i.e., according to the first voltage threshold Ul. Preferably, each electrical charge 50 is an electrical resistor having a resistance between 0.01 Q - Ohm - and 10 Q.

[0052] Each electrical load 50 comprises two terminals, each of which is connected to a bipolar plate 36 of the associated unit cell 32 via a respective switch 52. In practice, the switches 52 are preferably Semiconductor devices such as transistors or thyristors, MOSFETs, or equivalents. These elements can advantageously be integrated into the control device 40, as illustrated in [Fig. 2]. The electrical loads 50 serve both to limit the growth of the unit voltage of the cells, according to the principles of the invention described below, and to measure the impedance of each cell using the four-wire measurement method. When the two switches 52 associated with an electrical load 50 are closed, the corresponding unit cell 32 is connected to that electrical load 50. The switches 52 are selectively controllable, meaning that each switch 52 can be closed or opened independently of the other switches 52. Each electrical load 50 is thus selectively connectable, reversibly, to its associated unit cell 32. When a unit cell 32 is connected to the associated electrical load 50, this unit cell 32 is said to be in coupled configuration.

[0053] During the start-up of the fuel cell 20, all the unit cells 32 are simultaneously supplied with gas; however, the evolution of the unit voltage Uc of each unit cell 32 of the stack 30 is not perfectly synchronized, so that some unit cells 32 see their unit voltage Uc increase faster than other unit cells 32. [Fig. 3]#a) is a graph 100A representing the evolution of the unit voltage Uc of the unit cells 32 of the stack 30. Depending on the application, a stack 30 comprises up to several dozen unit cells, for example 60 or 80 unit cells.

[0054] On the 100A graph, only the unit voltage Uc of four unit cells 32 is shown to avoid cluttering the graph. In the illustrated example, the unit voltage Uc of each unit cell 32 is initially zero, equal to 0 V. From an initial instant T0, the unit voltage Uc of the first unit cell 32 begins to increase, until it reaches a unit voltage approximately equal to an operating voltage UF, generally on the order of 1 V. In the illustrated example, the operating voltage is approximately equal to 0.96 V. From a final instant TF, the unit voltage Uc of each of the unit cells 32 is considered to be approximately equal, and stable, to the operating voltage UF.

[0055] The time period between the initial instant T0 and the final instant TF corresponds to a phase of rising cell voltages of the unit cells 32 and lasts for example a few seconds, typically between 1 and 10.

[0056] Graph 100A thus includes a first curve 101, which corresponds to the unit cell 32, called the "first cell," whose unit voltage Uc increased earliest, compared to the other unit cells 32. Empirically, it has been observed that when one of the unit cells 32 saw its unit voltage Uc initially increase before the other unit cells 32, then generally this unit cell remained either the first, or at least among the first unit cells 32.

[0057] Thus, schematically and in a simplified manner, within the framework of the present description it is considered that for any two unit cells 32 of the stack 30, if, at any instant between the initial instant T0 and the final instant TF, the unit voltage Uc of one of these two unit cells 32 is greater than the unit voltage Uc of the other unit cell, then at any other instant between the initial instant T0 and the final instant TF, the unit voltage Uc of the first of these two unit cells 32 remains greater than the unit voltage Uc of the other unit cell.

[0058] Graph 100A also includes a second curve 102, which corresponds to the unit cell 32 - called the "last cell" - whose unit voltage Uc increased last, compared to the other unit cells 32. It is understood that the first and second curves 101 and 102 form an envelope for all the voltage curves of the unit cells 32 of the stack 30.

[0059] Graph 100A also includes a third curve 103, called the median curve, which divides the unit cells 21 into two groups of equal size. In other words, on graph 102, the first half of the unit cells 32 have a unit voltage Uc located above the median curve 103, while the second half of the unit cells 32 have a unit voltage Uc located below the median curve 103. Schematically, the dispersion of the growth of the unit voltage Uc during the rise phase is considered random, so the dispersion around the third curve 103 follows a roughly Gaussian distribution, and the third curve 103 is roughly located midway between the first and second curves 101 and 102.

[0060] Graph 100A also includes a fourth curve 104, which corresponds to a unit cell 32 whose unit voltage Uc has increased less rapidly than that of the first cell, while being located above the median curve 103.

[0061] Graph 100A also includes a fifth curve 105, which corresponds to a unit cell 32 whose unit voltage Uc has increased faster than that of the last, while being located below the median curve 103.

[0062] Fig. 3#b) is a graph 100B, which is a simplified version of graph 100A. On graph 100B, only the first curve 101, second curve 102 and median curve 103 are represented.

[0063] At any instant during the rise phase, a voltage difference AU between the cell voltage of the first cell and the unit voltage Uc of the last cell evolves, this voltage difference AU being zero at the initial instant T0, zero at the final instant TF, and passing through a maximum. In the example illustrated in graph 100B, the voltage difference AU reaches a maximum value, denoted AUmax, approximately equal to 0.55 V. It is understood that the further apart the first and second curves 101 and 102 are, the greater the maximum voltage difference AUmax, which increases the risk of corrosion in the stack 30, and in particular the risk of corrosion of the carbon contained in the catalytic layers of the membrane. This corrosion will then directly impact the lifespan and durability of the fuel cell.

[0064] A method for controlling the fuel cell 20 is thus implemented to reduce this risk of corrosion. This control method is explained with reference to Figures 4 and 5.

[0065] In a first step, for each unit cell 32, the unit voltage Uc is measured by means of the control device 40, so as to follow an evolution of the unit voltage Uc as a function of time.

[0066] Next, the control method includes an identification step, during which one or more unit cells 32, called fast cells, are identified, whose unit voltage Uc reaches a first predetermined voltage threshold Ul, the first threshold being non-zero and lower than the operating voltage UF.

[0067] The first cell is therefore the first whose unit voltage Uc reaches the first voltage threshold Ul. In the example of [Fig.4], the first curve 101 reaches the first voltage threshold Ul at a so-called "connection" instant Tl, which occurs for example a few milliseconds or seconds after the initial instant T0.

[0068] Preferably, the first voltage threshold Ul is between 0.05 V and 0.6 V, preferably between 0.05 V and 0.4 V, preferably also equal to 0.1 V ± 0.05 V. In the illustrated example, the first voltage threshold Ul is equal to 0.3 V.

[0069] Next, the control method includes a regulation step, during which at least one of the fast cells is put into coupled configuration, the at least one fast cell being connected to the corresponding electrical load 50, so as to limit the increase in the unit voltage Uc of the at least one fast cell considered.

[0070] As long as the unit cell 32 remains connected to the electrical load 50, in other words, remains in a coupled configuration, an electric current flows through this electrical load 50, resulting in the dissipation of electrical energy. This contributes to slowing the growth of the unit voltage Uc of this fast cell compared to a situation where this fast cell 32 would not have been in a coupled configuration. It is thus possible, thanks to the invention, to harmonize the time taken by the different cells to reach the operating voltage UF. Consequently, the risk of corrosion occurring within the cells is limited, if not eliminated, and the battery's durability is improved.

[0071] With reference to the schematic example in [Fig. 4], the first cell is put into coupled configuration from the moment of connection T1. The unit voltage Uc, instead of continuing along curve 101, it bifurcates and grows more slowly than in the absence of connection to the electric charge 50, the unit voltage Uc following a curve 101'. On [Fig.4], the bifurcation of curve 101 to curve 101' occurs at a first bifurcation point PI.

[0072] The first cell is obviously not left connected to the corresponding electrical load 50 indefinitely, because otherwise the unit voltage Uc of this first cell would become lower than the unit voltage of the last cell, which is undesirable. Furthermore, each unit cell 32 in the coupled configuration dissipates energy, which causes the fuel cell 20 to heat up. Thus, the first cell is advantageously disconnected from the corresponding electrical load 50 at a disconnection time T2, at the latest when the unit voltage Uc of the first cell is equal to the unit voltage Uc of the last cell. In the illustrated example, to disconnect the electrical load 50, it suffices to open at least one of the two switches 52 associated with this electrical load 50, the first cell switching from the coupled configuration to a decoupled configuration.Thus, thanks to the invention, and in particular to the rapid and individual disconnection of each unit cell 32 and the corresponding electrical charge 50, it is possible to limit the overall heating of the fuel cell 20, and thereby improve the safety of use and the durability of the cell. For example, thanks to the invention it is possible to dissipate only the electrical power of a few cells (for example 10% to 50% of the total number of cells) and therefore to dissipate as heat only this power (i.e., between 10% and 50% of the heat that would have been dissipated according to the prior art method).

[0073] Once the first cell is disconnected from the corresponding electric charge 50, in other words decoupled, the unit voltage Uc of the first cell begins to grow again at a rate similar to that of the other unit cells 32 which have remained decoupled, disconnected from any electric charge 50. On [Fig.4], the unit voltage Uc follows the curve 101' as long as the first cell is in coupled configuration, between times T1 and T2, then bifurcates from time T2, at a second bifurcation point P2, onto a curve 101” once the electric charge 50 is disconnected.

[0074] By temporarily connecting the first cell to the electrical load 50, the unit voltage Uc of the first cell is brought closer to the unit voltage Uc of the last cell, thus reducing the risk of corrosion. The same operation of connecting to the associated electrical load 50 is advantageously repeated for several high-speed cells.

[0075] For example, during the regulation step, as long as a proportion of the number of fast cells relative to the total number of unit cells 32 is less than a first predetermined percentage, then each fast cell is connected to the associated electrical load 50. According to examples, the first percentage is between 20% and 80%, preferably less than 70%, preferably even less than 60%.

[0076] Advantageously, once the proportion of fast cells connected to the corresponding electrical load 50 relative to the total number of unit cells 32 is equal to the first percentage, no further unit cells 32 are connected to the corresponding electrical load, in other words, no further unit cells 32 are put into coupled configuration.

[0077] As mentioned previously, the unit cells 32 in coupled configuration are not left indefinitely in that configuration. Several strategies—called decoupling strategies—are therefore available to determine the best time to decouple the unit cells 32.

[0078] According to a first decoupling strategy, each fast cell, once connected to the associated electrical load 50, is disconnected from the associated electrical load no later than the end of a first predetermined time interval after its connection to the corresponding electrical load 50. Preferably, the first time interval is between 0.05 s and 0.6 s. In particular, the upper limit of the first time interval is fixed, so as to ensure that all the individual cells 32 are decoupled at the end of the start-up phase.

[0079] According to a second decoupling strategy, all unit cells 32 in coupled configuration are decoupled after a second predetermined time interval after the first unit cell 32 has been put into coupled configuration, the second time interval being calculated based on measurements of the evolution of cell voltages.

[0080] For example, a rise time T3 is defined as the instant when half of the unit cells 32 reach a second voltage threshold U2, the second voltage threshold U2 being between 0.1 V and 0.6 V and being less than or equal to the first voltage threshold Ul, so that the mean rise time T3 is not influenced by the coupled configuration of the unit cells 32. Preferably, the second voltage threshold U2 is greater than 0.15 V, and even more preferably greater than 0.20 V. In the example illustrated in [Fig. 4]#b), the second voltage threshold U2 is advantageously chosen to be equal to the first voltage threshold Ul, so the rise time T3 corresponds to the intersection between the curve 103 and the first voltage threshold Ul.

[0081] An average delay ATup is also defined, which is equal to the time elapsed between the moment when the unit voltage Uc of the first fast cell reaches the second voltage threshold U2 and the moment when the unit voltage Uc of half of the unit cells 32 reaches the second voltage threshold U2. In the example of [Fig. 4]#a), the second Since the voltage threshold U2 is equal to the first voltage threshold Ul, the average delay ATup is equal to the time difference between the rise time T3 and the connection time T1.

[0082] Schematically, it can be estimated that the last unit cell 32 will reach the second voltage threshold U2 after a time interval approximately equal to twice the average delay ATup.

[0083] Advantageously, the second time interval is proportional to the average delay ATup, that is, equal to a coefficient a multiplied by the average delay ATup. The coefficient a is chosen to be strictly positive and less than 2, preferably greater than 1, and preferably still between 1.2 and 1.8.

[0084] Figure 5 is a graph that schematically represents an effect of the control method of the invention when the first percentage is approximately 60% and the first voltage threshold Ul is 0.3 V. As long as 60% of the unit cells 32 have not been connected to the corresponding electrical load 50, then as soon as the unit voltage Uc of another unit cell 32 reaches the first voltage threshold Ul, that unit cell 32 is put into a coupled configuration. Once 60% of the unit cells 32 have been put into the coupled configuration, then no other unit cells 32 are put into the coupled configuration. All the unit cells 32 that have been put into the coupled configuration are decoupled simultaneously at time T2.

[0085] In the example of [Fig.5], the maximum voltage difference AUmax between the cell voltages Uc of any two unit cells 32, during the start-up phase, is approximately equal to 0.25 V, that is to say much lower than the situation where the method of the invention is not implemented, as shown in [Fig.3], where the maximum voltage difference AUmax is approximately equal to 0.55 V. The risk of corrosion is thus reduced.

[0086] More generally, it is understood that various decoupling strategies can be implemented in order to optimize the reduction of the risk of corrosion by leaving the unit cells 32 in coupled configuration for as long as possible, while minimizing heating due to the consumption of electrical energy in the electrical loads 50.

[0087] In the illustrated example, the control module 40 is configured to measure the unit voltage Uc across the terminals of each unit cell 32.

[0088] According to an alternative not shown, the unit cells 32 are distributed into several subgroups of juxtaposed cells. Each subgroup of cells, also simply called a "subgroup", comprises at least one unit cell 32 and forms a subset of the stack 30. The situation where each subgroup comprises only one unit cell 32 corresponds to the case described above, the principles of the invention can be transposed to the case where each subgroup comprises several unit cells 32.

[0089] Preferably, each cell subgroup comprises several adjacent unit cells 32. Preferably, for any two given cell subgroups, the unit cells of one subgroup are distinct from the unit cells of the other subgroup. Preferably, all cell subgroups comprise the same number of unit cells 32. Preferably, a cell subgroup comprises between ten and thirty unit cells 32, advantageously twenty unit cells 32. For example, the stack 30 comprises a total of two hundred unit cells 32, which are grouped into ten cell subgroups, each comprising twenty unit cells 32.

[0090] The control module 40 is advantageously configured to measure an electrical voltage across each subgroup of cells. The electrical voltage measured across each subgroup of cells is also called the unit voltage. The unit voltage of each subgroup is equal to the sum of the unit voltage Uc of each unit cell 32 that makes up the subgroup in question. Similarly, an operating voltage of each subgroup is equal to the sum of the operating voltage UF of each unit cell 32 that makes up the subgroup in question.

[0091] With reference to [Fig. 6], the method for controlling the fuel cell 20 includes a start-up phase 601 of the cell, during which the unit voltage of each subgroup is initially zero and then increases until it reaches an electrical voltage substantially equal to the operating voltage of the subgroup in question. Each cell subgroup is initially decoupled from the associated electrical load.

[0092] During the start-up phase, the method comprises, for each subgroup, the measurement 602 of the associated unit voltage Uc by means of the control device 40, so as to monitor the evolution of the unit voltage of the subgroup in question as a function of time. The measurement of the unit voltage Uc is repeated several times during the start-up phase 601, preferably periodically, for example, 1000 times during the start-up phase. By way of illustration, if the start-up phase lasts approximately 10 seconds, then the measurement of the unit voltage Uc is repeated every 10 ms.

[0093] The process also includes an identification step 603, during which one or more subgroups, referred to as fast subgroups, are identified, whose unit voltage is the first to reach a predetermined first voltage threshold, the first voltage threshold being non-zero and lower than the operating voltage of the subgroup in question. For each cell subgroup, if the result of the identification step 603 is negative, as represented by the letter N, so the subgroup in question is not a fast subgroup and remains decoupled from the associated electric charge. We return to the previous step of measuring the unit voltage Uc.

[0094] If the result of the identification step 603 is positive, as represented by the letter Y, then the subgroup considered is a fast subgroup.

[0095] For each fast subgroup, the control method includes a regulation step 604, during which at least one of the fast subgroups is put into a coupled configuration, in which the at least one subgroup is connected to a respective electrical load, so as to limit the increase in the unit voltage of the at least one subgroup considered.

[0096] As mentioned previously, once placed in a coupled configuration, each fast subgroup does not remain indefinitely in the coupled configuration. The control method includes a test phase 605, during which it is determined whether the subgroup in the coupled configuration should be placed in a decoupled configuration or left in the coupled configuration. For each subgroup in the coupled configuration, the test phase is advantageously repeated several times, preferably periodically, for example every 10 ms. One or more of the decoupling strategies are thus implemented during the test phase 605.

[0097] For each subgroup of cells in coupled configuration, if the result of test phase 605 is negative, as represented by the letter N, then the subgroup in question is left in the coupled configuration with the associated electrical load. This returns to the previous regulation step 604. If the result of test phase 605 is positive, as represented by the letter Y, then during a decoupling step 606 subsequent to test phase 605, the subgroup in question is decoupled from the corresponding electrical load.

[0098] The embodiments and variants mentioned above can be combined with each other to generate new embodiments of the invention.

Claims

Demands

1. A method for controlling a fuel cell (20), the fuel cell comprising: • a stacking (30) of several unit cells (32), the unit cells being distributed into several juxtaposed subgroups, each subgroup comprising at least one unit cell, • a control device (40), which is configured to measure a unit voltage (Uc), which is equal to an electrical voltage across the terminals of each subgroup, in which, during a fuel cell start-up phase, while the unit electrical voltage (Uc) of each subgroup is initially zero and then increases until it reaches an electrical voltage substantially equal to an operating voltage (UF), the method comprises: a. For each subgroup, the measurement of the associated unit voltage (Uc) using the control device (40), so as to monitor the evolution of the unit voltage as a function of time, b. an identification step, during which one or more of the subgroups, called fast subgroups, are identified, whose unit voltage (Uc) first reaches a predetermined first voltage threshold (Ul), the first voltage threshold (Ul) being non-zero and lower than the operating voltage (UF), c. a regulation step, during which at least one of the fast subgroups is put into a coupled configuration, in which the subgroup in question is connected to a respective electrical load (50), so as to limit the increase in the unit voltage (Uc) of the subgroup in question.

2. A piloting method according to claim 1, wherein: • during the regulation stage, as long as a proportion of the number of fast subgroups relative to the total number of subgroups in the stack (30) is less than a first predetermined percentage, then each sub- fast group is connected to the associated electrical load (50).

3. A piloting method according to claim 2, wherein: • the first percentage is between 20% and 80%, preferably less than 70%, preferably still less than 60%.

4. A piloting method according to any one of claims 2 or 3, wherein: • once the proportion of fast subgroups connected to the corresponding electrical load (50) is equal to the first percentage, no further subgroups are subsequently connected to the corresponding electrical load.

5. A piloting method according to any one of claims 1 to 4, wherein each fast subgroup, once connected to the associated electrical load (50), is disconnected from the associated electrical load no later than after a first predetermined time interval.

6. A control method according to claim 5, wherein: • an average voltage rise delay (ATup) is defined, which is equal to the time elapsed between the moment when the unit voltage (Uc) of the first fast subgroup reaches a second voltage threshold (U2) and the moment when the unit voltage (Uc) of half of the subgroups reaches the second voltage threshold (U2), the second voltage threshold (U2) being less than or equal to the first voltage threshold (Ul), • the first time interval is calculated as a function of the average voltage rise delay (ATup).

7. A control method according to claim 6, wherein the second voltage threshold (U2) is equal to the first voltage threshold (Ul).

8. A piloting method according to any one of claims 6 or 7, wherein the first time interval is equal to a coefficient (a) multiplied by the average delay (ATup), the coefficient (a) being less than 2, preferably greater than 1, preferably still between 1.2 and 1.

8.

9. A control method according to any one of the preceding claims, wherein: • the first voltage threshold (Ul) is between 0.05 V and 0.6 V, preferably between 0.05 V and 0.4 V, preferably also equal to 0.1 V ± 0.05 V.

10. Fuel cell (20), comprising: • a stack (30) of several unit cells (32), the unit cells being distributed into several juxtaposed subgroups, each subgroup comprising at least one unit cell (32), • a control device (40), which is configured to measure a unit voltage (Uc) of each subgroup, and which comprises several electrical loads (50), each electrical load being associated with a respective subgroup being selectively connectable, reversibly, to the associated subgroup, wherein the fuel cell (20) is configured to implement the control method according to any one of the preceding claims.

11. Fuel cell (20) according to claim 10, wherein: • the control device (40) also includes a switching device between each subgroup and each associated electrical load (50).

12. Fuel cell (20) according to claim 11, wherein: • each switching device is a switch.

13. Fuel cell (20) according to any one of claims 10 to 12, wherein: • each electrical charge (50) is a resistor.

14. Fuel cell (20) according to claim 13, wherein: • the control module is configured to perform an impedance measurement across the terminals of the corresponding subgroup according to a four-wire method and using the corresponding resistance.

15. Vehicle (10), comprising a fuel cell (20) according to any one of claims 10 to 14.

Images