Fuel cell system

Abstract

This fuel cell system includes: a first supply unit that has a first injector and a first ejector and supplies a fuel gas to a fuel cell via a first supply flow-channel connected to a gas inflow unit; a second supply unit that has a second injector and a second ejector and supplies a fuel gas to the fuel cell via a second supply flow-channel connected to the gas inflow unit; and a control unit. A circulation flow ratio of the second ejector is higher than the circulation flow ratio of the first ejector. When the first injector and the second injector inject the fuel gas, if required power increases, the control unit controls the first injector and the second injector such that an average value of an injection amount per unit time increases more in the first injector than in the second injector, while if the required power decreases, the control unit controls the first injector and the second injector such that the average value of the injection amount per unit time decreases more in the first injector than in the second injector.

Description

Technical Field

[0005] ,

[0001] The present invention relates to a fuel cell system having a plurality of ejectors.

Background Art

[0002] In recent years, efforts to achieve a low-carbon society or a decarbonized society have been active, and research and development on fuel cells have been conducted in vehicles in order to reduce CO2 emissions and improve energy efficiency. In this regard, conventionally, a fuel cell system has been known in which a pair of ejectors having different circulation capabilities are provided in a flow path for supplying fuel gas to a fuel cell (see, for example, Patent Document 1). In the system described in Patent Document 1, when the required power is lower than a threshold value, fuel gas is supplied to the fuel cell through a first ejector having a high circulation capacity. When the discharge of the fuel gas from the first ejector becomes abnormal, the supply from the first ejector is stopped, and fuel gas is supplied to the fuel cell through a second ejector having a low circulation capacity.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in a fuel cell system having a pair of ejectors with different circulation capabilities, there is room for improvement in the system described in Patent Document 1 from the viewpoint of optimizing the drainage from the fuel cell, particularly during transient operation.

Means for Solving the Problems

[0005] A fuel cell system according to one aspect of the present invention comprises a fuel cell that generates electricity by an electrochemical reaction between a fuel gas and an oxidizer gas and has a gas inlet into which fuel gas flows and a gas outlet into which fuel gas flows out; a first supply unit having a first injector for injecting fuel gas and supplying fuel gas to the fuel cell via a first supply channel connected to the gas inlet; a second supply unit having a second injector for injecting fuel gas and supplying fuel gas to the fuel cell via a second supply channel connected to the gas inlet; and a control unit that controls the first injector and the second injector to inject fuel gas in an injection mode according to the required power. The first supply unit further has a first ejector provided between the first injector and the gas inlet that guides fuel gas flowing out from the gas outlet to the first supply channel via a first circulation channel, and the second supply unit further has a second ejector provided between the second injector and the gas inlet that guides fuel gas flowing out from the gas outlet to the second supply channel via a second circulation channel. The first injector and the second injector are configured such that the amount of fuel gas injected per unit time by the first injector is greater than the amount of fuel gas injected per unit time by the second injector. The first ejector and the second ejector are configured such that the first circulation flow rate ratio, which is the ratio of the circulation amount by the first ejector to the amount of fuel gas injected by the first injector, is smaller than the second circulation flow rate ratio, which is the ratio of the circulation amount by the second ejector to the amount of fuel gas injected by the second injector. When the injection mode is a predetermined injection mode in which the first injector and the second injector each inject fuel gas, the control unit controls the first injector and the second injector such that when the required power increases, the average value of the injection amount per unit time increases for the first injector more than for the second injector, and when the required power decreases, the average value of the injection amount per unit time decreases for the first injector more than for the second injector. [Effects of the Invention]

[0006] According to the present invention, it is possible to improve the drainage performance from the fuel cell stack during transient operation. [Brief explanation of the drawing]

[0007] [Figure 1] A schematic perspective view showing the overall configuration of a fuel cell stack according to an embodiment of the present invention. [Figure 2] Figure 1 is a perspective view showing the schematic configuration of the integrated electrode assembly included in the fuel cell stack. [Figure 3] Figure 1 illustrates the accumulation of water within the fuel cell stack. [Figure 4] A schematic diagram showing the main components of a fuel cell system according to an embodiment of the present invention. [Figure 5] Figure 4 shows the configuration of the ejector included in the fuel cell system. [Figure 6] A diagram showing the relationship between the amount of fuel injected by the injector and the amount of fuel circulated by the ejector. [Figure 7A] Figure 4 shows an example of a fuel gas injection pattern from a fuel cell system. [Figure 7B] This figure shows a modified version of Figure 7A. [Figure 8] This figure shows an example of how the differential pressure between the inlet and outlet of a fuel cell stack and the drainage rate of stagnant water change over time, obtained by a fuel cell system according to an embodiment of the present invention. [Figure 9] Figure 8 shows a comparative example. [Figure 10] This diagram shows the relationship between the delay time from when the injector on the large ejector side finishes spraying until the injector on the small ejector side starts spraying, and the drainage rate of stagnant water. [Figure 11] A flowchart showing an example of the process performed by the ECU in Figure 4. [Figure 12] Figure 4 shows the relationship between the injection cycle of the injector and the valve opening time. [Figure 13] Figure 1 shows the stable and unstable regions of power generation in a fuel cell stack, depending on the instantaneous gas flow velocity and average gas flow rate. [Figure 14A] A time chart showing an example of operation during transient operation of a fuel cell system according to an embodiment of the present invention. [Figure 14B]A time chart illustrating another example of the operation of a fuel cell system according to an embodiment of the present invention during transient operation. [Figure 15A] This flowchart shows an example of the processing performed by the ECU in Figure 4, and is a flowchart showing an example of the processing corresponding to the time chart in Figure 14A. [Figure 15B] This flowchart shows an example of the processing performed by the ECU in Figure 4, and is a flowchart showing an example of the processing corresponding to the time chart in Figure 14B. [Figure 16] This figure shows an example of how the injector's injection pattern changes when the power requirement increases. [Modes for carrying out the invention]

[0008] Embodiments of the present invention will be described below with reference to Figures 1 to 16. The fuel cell system according to the embodiment of the present invention has a fuel cell stack, which is the main body of the fuel cell. The fuel cell is mounted on a vehicle, for example, and generates electricity for driving the vehicle. First, the overall configuration of the fuel cell stack will be described in general terms. Note that the fuel cell stack may be simply called a fuel cell, or it may be called the fuel cell body.

[0009] Figure 1 is a schematic perspective view showing the overall configuration of a fuel cell stack 100 according to an embodiment of the present invention. For convenience, the three mutually orthogonal axial directions shown in the figure will be defined as the longitudinal direction, the left-right direction, and the vertical direction, and the configuration of each part will be described according to this definition. These directions are not necessarily the same as the longitudinal, left-right, and vertical directions of a vehicle. For example, the longitudinal direction in Figure 1 may be the longitudinal direction of a vehicle, the left-right direction, or the vertical direction.

[0010] As shown in FIG. 1, the fuel cell stack 100 includes a cell stack 101 formed by stacking a plurality of power generation cells 1 in the front-back direction, and end units 102 disposed at both front and rear ends of the cell stack 101, and has an overall substantially rectangular parallelepiped shape. The length of the cell stack 101 in the left-right direction is longer than the length in the up-down direction. For convenience, a single power generation cell 1 is shown in FIG. 1. The power generation cell 1 includes an integrated electrode assembly 2 (UEA; Unitized Electrode Assembly) having a membrane electrode assembly including an electrolyte membrane and electrodes, and separators 3, 3 disposed on both front and rear sides of the integrated electrode assembly 2 to sandwich the integrated electrode assembly 2. The integrated electrode assembly 2 and the separators 3 are alternately arranged in the front-back direction. The integrated electrode assembly 2 may be referred to as a membrane electrode structure.

[0011] The separator 3 has a pair of front and rear thin metal plates with a corrugated cross-section, and the outer peripheral edges of these thin plates are joined together to form an integral structure. A conductive material with excellent corrosion resistance is used for the separator 3, and for example, titanium, titanium alloy, stainless steel, etc. can be used. Inside the separator 3 (between the pair of thin plates), a cooling flow path through which a cooling medium flows is formed, and the power generation surface of the power generation cell 1 is cooled by the flow of the cooling medium. For example, water can be used as the cooling medium. A mixed solution of water and ethylene glycol can also be used. The surfaces (front surface and rear surface) of the separator 3 facing the integrated electrode assembly 2 are formed in a concavo-convex shape by press molding or the like so as to form a gas flow path between the separator 3 and the membrane electrode assembly of the integrated electrode assembly 2. The cooling flow path can also be formed by a mesh flow path, or can be formed by a refrigerant surface welding type in which a member is welded to the refrigerant surface to form the flow path.

[0012] The separator 3 on the front side of the integrated electrode assembly 2 is, for example, an anode-side separator (anode separator), and an anode flow path through which fuel gas flows is formed between the anode separator 3 and the membrane electrode assembly of the integrated electrode assembly 2. The separator 3 on the rear side of the integrated electrode assembly 2 is, for example, a cathode-side separator (cathode separator), and a cathode flow path through which oxidant gas flows is formed between the cathode separator 3 and the membrane electrode assembly of the integrated electrode assembly 2. For example, hydrogen gas can be used as the fuel gas, and air can be used as the oxidant gas.

[0013] FIG. 2 is a perspective view showing a schematic configuration of the integrated electrode assembly 2. As shown in FIG. 2, the integrated electrode assembly 2 includes a substantially rectangular membrane electrode assembly 20 (MEA; Membrane Electrode Assembly) and a frame 21 that supports the membrane electrode assembly 20. The membrane electrode assembly 20 includes an electrolyte membrane, an anode electrode provided on the front surface of the electrolyte membrane, and a cathode electrode provided on the rear surface of the electrolyte membrane.

[0014] The electrolyte membrane is, for example, a solid polymer electrolyte membrane, and a thin film of a perfluorosulfonic acid polymer containing moisture can be used. Not limited to fluorine-based electrolyte membranes, hydrocarbon-based electrolyte membranes can also be used.

[0015] The anode electrode is formed on the front surface of the electrolyte membrane and includes an electrode catalyst layer that serves as a reaction field for the electrode reaction, and a gas diffusion layer provided on the front surface of the electrode catalyst layer that diffuses and supplies fuel gas. The cathode electrode is formed on the rear surface of the electrolyte membrane and includes an electrode catalyst layer that serves as a reaction field for the electrode reaction, and a gas diffusion layer provided on the rear surface of the electrode catalyst layer that diffuses and supplies oxidant gas. The electrode catalyst layer contains a catalyst metal that promotes the electrochemical reaction between hydrogen contained in the fuel gas and oxygen contained in the oxidant gas, an electrolyte having proton conductivity (such as an ionomer), and carbon particles having electron conductivity. The gas diffusion layer is composed of a conductive member having gas permeability, such as a carbon porous body.

[0016] At the anode electrode, fuel gas (hydrogen) supplied via the anode channel and gas diffusion layer is ionized by the action of a catalyst and moves to the cathode electrode side through the electrolyte membrane. The electrons generated at this time pass through an external circuit and are extracted as electrical energy. At the cathode electrode, oxidizing gas (oxygen) supplied via the cathode channel and gas diffusion layer reacts with hydrogen ions introduced from the anode electrode and electrons that have moved from the anode electrode to produce water. The generated water provides appropriate humidity to the electrolyte membrane, and excess water is discharged to the outside of the integrated electrode assembly 2 along the gas flow.

[0017] The frame 21 is a thin plate with a roughly rectangular shape and is made of an insulating resin or rubber. A roughly rectangular opening 21a is provided in the center of the frame 21. The membrane electrode assembly 20 is positioned to cover the entire opening 21a and is supported by the periphery of the opening 21a. Three through holes 211 to 213 are opened vertically on the left side of the opening 21a, penetrating the frame 21 in the front-to-back direction, and three through holes 214 to 216 are opened vertically on the right side of the opening 21a, penetrating the frame 21 in the front-to-back direction.

[0018] As shown in Figure 1, the front and rear separators 3 of the integrated electrode assembly 2 have through-holes 311 to 316 that penetrate the separator 3 in the front-to-back direction, at positions corresponding to the through-holes 211 to 216 of the frame 21. The through-holes 311 to 316 communicate with the through-holes 211 to 216 of the frame 21, respectively. The collection of these interconnected through-holes 211 to 216 and 311 to 316 forms flow channels PA1 to PA6 (indicated by arrows for convenience) that penetrate the cell stack 101 and extend in the front-to-back direction. Flow channels PA1 to PA6 are sometimes called manifolds. Flow channels PA1 to PA6 are connected to a manifold outside the fuel cell stack 100.

[0019] The flow path PA1 (solid arrow) extending forward through through holes 211 and 311 is the fuel gas supply flow path. The flow path PA6 (solid arrow) extending backward through through holes 216 and 316 is the fuel gas discharge flow path. The fuel gas supply flow path PA1 and the fuel gas discharge flow path PA6 communicate with the anode flow path, which is located opposite the front surface of the membrane electrode assembly 20. As shown by the solid arrows, fuel gas flows from left to right through the anode flow path via the fuel gas supply flow path PA1 and the fuel gas discharge flow path PA6. Communication between the anode flow path and the other flow paths PA2 to PA5 is blocked via a seal (not shown). The fuel gas flowing through the fuel gas discharge flow path PA6 is the fuel gas after a portion has been used at the anode electrode, and this is sometimes called fuel exhaust gas.

[0020] The flow path PA4 (dotted arrow) extending forward through through holes 214 and 314 is the oxidizer gas supply flow path. The flow path PA3 (dotted arrow) extending backward through through holes 213 and 313 is the oxidizer gas discharge flow path. The oxidizer gas supply flow path PA4 and the oxidizer gas discharge flow path PA3 communicate with the cathode flow path provided opposite the rear surface of the membrane electrode assembly 20, and as shown by the dotted arrow, the oxidizer gas flows from right to left through the cathode flow path via the oxidizer gas supply flow path PA4 and the oxidizer gas discharge flow path PA3. Communication between the cathode flow path and the other flow paths PA1, PA2, PA5, PA6 is blocked via a seal portion (not shown). The oxidizer gas flowing through the oxidizer gas discharge flow path PA3 is the oxidizer gas after a portion has been used at the cathode electrode, and this is sometimes called oxidizer exhaust gas. Sometimes, fuel exhaust gas and oxidizer exhaust gas are not distinguished and are collectively called reaction exhaust gas.

[0021] The flow path PA5 (dotted arrow) extending forward through through holes 215 and 315 is a cooling medium supply flow path. The flow path PA2 (dotted arrow) extending backward through through holes 212 and 312 is a cooling medium discharge flow path. The cooling medium supply flow path PA5 and the cooling medium discharge flow path PA2 are in communication with the cooling flow path inside the separator 3, and the cooling medium flows through the cooling flow path via the cooling medium supply flow path PA5 and the cooling medium discharge flow path PA2. Communication between the cooling flow path and the other flow paths PA1, PA3, PA4, PA6 is blocked via a seal portion (not shown).

[0022] The end units 102, located on both the front and rear sides of the cell stack 101, each have a terminal plate 4, an insulating plate 5, and an end plate 6. The rear end unit 102 has multiple through holes 102a to 102f that penetrate the end unit 102 in the front-rear direction. Through hole 102a opens on the extension of the fuel gas supply channel PA1 and communicates with the fuel gas supply channel PA1. Through hole 102b opens on the extension of the cooling medium discharge channel PA2 and communicates with the cooling medium discharge channel PA2. Through hole 102c opens on the extension of the oxidizer gas discharge channel PA3 and communicates with the oxidizer gas discharge channel PA3. Through hole 102d opens on the extension of the oxidizer gas supply channel PA4 and communicates with the oxidizer gas supply channel PA4. Through hole 102e opens on the extension of the cooling medium supply channel PA5 and communicates with the cooling medium supply channel PA5. The through-hole 102f opens on the extension of the fuel gas exhaust passage PA6 and communicates with the fuel gas exhaust passage PA6.

[0023] More specifically, a fuel gas tank containing high-pressure fuel gas is connected to the through-hole 102a via an ejector, injector, etc., as described later, and fuel gas is supplied to the fuel cell stack 100 through the through-hole 102a. Fuel gas (fuel exhaust gas) is discharged from the through-hole 102f. A compressor for supplying oxidizer gas is connected to the through-hole 102d, and compressed oxidizer gas is supplied to the fuel cell stack 100 through the through-hole 102d. Oxidizer gas (oxidizer exhaust gas) is discharged from the through-hole 102c. A pump for supplying cooling medium is connected to the through-hole 102e, and cooling medium is supplied to the fuel cell stack 100 through the through-hole 102e. Cooling medium is discharged from the through-hole 102b. The discharged cooling medium is cooled by heat exchange in the radiator and supplied back to the fuel cell stack 100 through the through-hole 102e.

[0024] The above is a general overview of the fuel cell stack 100. However, in the fuel cell stack 100, the water generated by power generation may remain near the membrane electrode assembly 20 and become stagnant water. Figure 3 is a schematic cross-sectional view of the main part of the membrane electrode assembly 20 illustrating the generation of such stagnant water. As shown by arrow A in Figure 3, the water generated at the cathode electrode flows through the electrolyte membrane 22 of the membrane electrode assembly 20 to the anode electrode and remains near the anode electrode, including the catalyst electrode layer 23 and the gas diffusion layer 24 (for example, in the region where the ionomer exists). For convenience, this water remaining in the membrane electrode assembly 20 is called MEA stagnant water W1.

[0025] Furthermore, even if the generated water that flows to the anode electrode flows into the anode channel 33 of the separator 3, some of the generated water and some of the condensation water generated in the anode channel 33 may accumulate in the anode channel 33. For convenience, this water that remains in the channel is called channel stagnant water W2. The generated water (MEA stagnant water W1, channel stagnant water W2) that accumulates in the fuel cell stack can cause power generation instability and electrode performance degradation, so it needs to be discharged promptly. Taking this into consideration, the fuel cell system in this embodiment is configured as follows.

[0026] Figure 4 is a schematic diagram showing the main components of a fuel cell system 10 according to an embodiment of the present invention. Figure 4 mainly shows the configuration related to the flow of fuel gas. As shown in Figure 4, the fuel cell system 10 has a tank 11 in which high-pressure fuel gas is stored, and a supply channel PA1 that connects the through-hole 102a, which is the fuel gas inlet, to the tank 11. The supply channel PA1 has a pair of supply channels PA11 and PA12 arranged in parallel with each other. A large flow rate of fuel gas can pass through the supply channel PA11. For convenience, the supply channel PA11 is called the large channel, and the supply channel PA12 is called the small channel. The fuel cell system 10 has a gas supply unit SP1 that supplies fuel gas to the fuel cell stack 100 via the large channel PA11, and a gas supply unit SP2 that supplies fuel gas to the fuel cell stack 100 via the small channel PA12.

[0027] The gas supply unit SP1 has an injector (INJ) 31 and an ejector (EJE) 41 interposed between the tank 11 and the through hole 102a on the large flow path PA11. The gas supply unit SP2 has an injector (INJ) 32 and an ejector (EJE) 42 interposed between the tank 11 and the through hole 102a on the small flow path PA12. Although not shown in the figures, there is a pressure reduction mechanism between the tank 11 and the injectors 31 and 32, and the gas supply units SP1 and SP2 each function as tank pressure supply mechanisms. The injector 31 is composed of multiple (3 in the figure) injectors (INJ) 31A, 31B, and 31C interposed in the flow paths PA111, PA112, and PA113, which are branched from the large flow path PA11 and provided in parallel.

[0028] The configuration of injector 32 and each of the injectors 31A to 31C (such as the diameter of the injection hole) are identical to each other. Injector 32 has a valve body that opens and closes the flow path PA12 and a coil that drives the valve body. Injectors 31A, 31B, and 31C each have a valve body that opens and closes the flow paths PA111, PA112, and PA113, and a coil that drives the valve body. Current flows through the coils of these injectors 31A to 31C and 32 in response to a command from the ECU 50, which drives the injectors 31A to 31C and 32 to open and close. That is, when injectors 31A to 31C and 32 are turned on by a command from the ECU 50, the coils are energized and the valves of injectors 31A to 31C and 32 open. When injectors 31A~31C and 32 are turned off by a command from ECU50, the power to the coils is cut off, and injectors 31A~31C and 32 close. Drive current is supplied to the coils of injectors 31A~31C and 32 via a driver circuit (not shown).

[0029] Injectors 31 and 32 are switched on and off by PWM control from the ECU 50. The ECU 50 controls the duty cycle (Ti / T0), which is the ratio of the valve opening time (pulse width) Ti to the pulse waveform period T0, thereby adjusting the amount of fuel gas injected by each valve opening of injectors 31A-31C and 32. Since injectors 31A-31C are installed in parallel, the maximum injection amount of injector 31 in the large flow path PA11 is greater than the maximum injection amount of injector 32 in the small flow path PA12. Hereafter, injector 31 may be referred to as the large injector and injector 32 as the small injector. The large injector 31 refers to the injector on the ejector 41 side, and the small injector 32 refers to the injector on the ejector 42 side.

[0030] The through-hole 102f (Figure 1), which is the outlet for the fuel gas, is connected to a gas-liquid separator 43. In the gas-liquid separator 43, water contained in the fuel exhaust gas is separated from the fuel exhaust gas and stored. A drain valve 44 is connected to the gas-liquid separator 43. When the drain valve 44 is opened, the water stored in the gas-liquid separator 43 is discharged to the outside. The drain valve 44 is opened and closed by a command from the ECU 50. The gas-liquid separator 43 is connected to ejectors 41 and 42, respectively, via circulation passages PA21 and PA22. The fuel exhaust gas separated in the gas-liquid separator 43 is recirculated to the ejectors 41 and 42 via circulation passages PA21 and PA22.

[0031] Figure 5 shows the schematic configuration of the ejector 41 (referred to as the large ejector for convenience) in the large flow path PA11. As shown in Figure 5, the large ejector 41 has a nozzle section 411, a suction section 412, a confluence section 413, and a diffuser section 414. The fuel gas injected from the injector 31 passes through the small-diameter nozzle section 411, as shown by the solid arrows in Figure 5, and then flows into the diffuser section 414 via the confluence section 413. At this time, the large ejector 41 draws in fuel exhaust gas from the circulation flow path PA21 via the suction section 412, as shown by the dotted arrows in Figure 5. The drawn-in fuel exhaust gas merges with the fuel gas that has passed through the nozzle section 411 at the confluence section 413, and after becoming a uniform flow in the diffuser section 414, it is supplied to the fuel cell stack 100 through the through-hole 102a (Figure 4).

[0032] Although not shown in the diagram, the ejector 42 in the small channel PA12 (referred to as the small ejector for convenience) also has a nozzle section, a suction section, a confluence section, and a diffuser section, similar to the large ejector 41. The fuel gas flow passing through the nozzle section draws fuel exhaust gas from the circulation channel PA22 via the suction section and supplies it to the fuel cell stack 100 through the through-hole 102a. The diameter of the nozzle section of the small ejector 42 is smaller than the diameter of the nozzle section 411 of the large ejector 41.

[0033] Figure 6 shows the relationship between the valve opening times Ti1 and Ti2, the injection amount, and the circulation amount when each injector 31 and 32 injects once. As shown in Figure 6, if we define the instantaneous injection amounts q10 and q20, obtained by dividing the injection amount Q0 from one injection of injectors 31 and 32 by the valve opening times Ti1 and Ti2, as the injection amount of injectors 31 and 32 per unit time, then q10 > q20.

[0034] Therefore, if we define the gas circulation rate per unit time by the large ejector 41 as q11 and the gas circulation rate per unit time by the small ejector 42 as q21, the circulation flow rate ratio, which is the ratio of the circulation rate (suction rate) q11 of the large ejector 41 to the injection rate q10 of the large injector 31 per unit time, is smaller than the circulation flow rate ratio, which is the ratio of the circulation rate (suction rate) q21 of the small ejector 42 to the injection rate q20 of the small injector 32 per unit time. Thus, the small ejector 42 has a higher circulation capacity (circulation performance, suction performance). Also, if we define the circulation rates by the large ejector 41 and the small ejector 42 in a single injection as Q11 and Q21, respectively, the fuel gas flow rate downstream of the small ejector 42 (Q0 + Q21) is greater than the fuel gas flow rate downstream of the large ejector 41 (Q0 + Q11).

[0035] The ECU 50 in Figure 4 is a computer comprising a processing unit having a CPU, ROM, RAM, and peripheral circuits. The ECU 50 receives a signal from an accelerator pedal position sensor 51 that detects the opening degree of the accelerator pedal (accelerator position). Based on the signal from the accelerator pedal position sensor 51, the ECU 50 calculates the power required to be generated by the fuel cell stack 100, i.e., the power required by the drive motor. The drive motor is an example of a load mounted on the vehicle. The ECU 50 may also calculate the total power required by considering the power required by other loads.

[0036] The power requirement calculated by the ECU 50 increases with increasing accelerator opening. The ECU 50 controls the injectors 31 and 32 to inject fuel gas in an injection mode corresponding to the power requirement. The injection modes include a dual injection mode in which fuel is injected from both injectors 31 and 32 in one cycle T0, and a single injection mode in which fuel is injected from only one injector 31 or 32. If we define a low-load state as when the power requirement (e.g., current density) is less than or equal to a predetermined value E1, a medium-load state as when the power requirement is greater than the predetermined value E1 and less than or equal to a predetermined value E2 which is greater than the predetermined value E1, and a high-load state as when the power requirement is greater than the predetermined value E2, the ECU 50 drives the injectors 31 and 32 on and off to inject fuel gas in both injection modes in all low-load, medium-load, and high-load states. That is, it controls the duty cycle of the large injector 31 and the duty cycle of the small injector 32 according to the power requirement. Furthermore, under predetermined load conditions (e.g., high load conditions), fuel gas may be injected in single-injection mode.

[0037] The duty cycle control (PWM control) of the large injector 31 includes sequentially selecting two injectors 31A, 31B, 31A, 31C, or 31B, 31C from the three injectors 31A to 31C, and turning the selected injectors 31A, 31B, 31A, 31C, or 31B, 31C on and off at the same timing. For example, first, injectors 31A and 31B are turned on and off at the same timing, then in the next cycle, injectors 31A and 31C are turned on and off at the same timing, and so on. When under low load conditions, one of the injectors 31A to 31C may be selected sequentially, and the selected injectors 31A to 31C may be turned on and off sequentially at predetermined timings. When under high load conditions, all three injectors 31A to 31C may be turned on and off at the same timing.

[0038] The ECU 50 may not always control the injectors 31 and 32 in both injection modes, but may control them in one injection mode depending on the power requirements. For example, it may switch only the large injector 31 on and off under low load conditions, and only the small injector 32 on and off under high load conditions. In other words, the ECU 50 may change the injection mode depending on the power requirements.

[0039] Figure 7A shows an example of the fuel gas injection pattern from injectors 31 and 32 in the fuel cell system 10 according to this embodiment. The horizontal axis of Figure 7A represents time, and the vertical axis represents the fuel gas flow rate q1 downstream of the large ejector 41 (q10 + q11 in Figure 6) and the fuel gas flow rate q2 downstream of the small ejector 42 (q20 + q21 in Figure 6). In the example shown in Figure 7A, at period T0, the large injector 31 is turned on for a predetermined valve opening time Ti1. Simultaneously with the turning on of the large injector 31, the small injector 32 is turned on for a predetermined valve opening time Ti2.

[0040] Since the instantaneous injection amount of the large injector 31, which is the injection amount per unit time, is greater than the instantaneous injection amount of the small injector 32 (Figure 6), the instantaneous gas flow rate q1 downstream of the large ejector 41 when the large injector 31 is turned on is greater than the instantaneous gas flow rate q2 downstream of the small ejector 42 when the small injector 32 is turned on. On the other hand, since the valve opening time (on period) Ti2 of the small injector 32 is longer than the valve opening time (on period) Ti1 of the large injector 31, the gas flow rate downstream of the small ejector 42 due to one valve opening of the small injector 32 is greater than the gas flow rate downstream of the large ejector 41 due to one valve opening of the large injector 31.

[0041] When the injector 31 is turned on at time ta in Figure 7A, the flow rate of the fuel gas flowing through the anode channel 33 (Figure 3) immediately becomes maximum, and the fuel gas flow velocity becomes maximum. A higher instantaneous gas flow rate enhances the drainage effect of the MEA stagnant water W1, thereby guiding the MEA stagnant water W1 into the anode channel 33.

[0042] In other words, the drainage effect of MEA stagnant water W1 is greater the faster the fuel gas flow velocity in the anode channel 33. In this embodiment, although the valve opening time Ti1 is short, a large flow rate q1 of fuel gas is supplied via the large ejector 41 at a predetermined period T0. Therefore, the MEA stagnant water W1 can be easily guided into the anode channel 33, and the MEA stagnant water W1 can be effectively discharged. In particular, in the example of Figure 7A, the small injector 32 turns on when the large injector 31 is turned on. Therefore, the instantaneous flow rate per unit time of fuel gas flowing through the anode channel 33 increases (q1 + q2) compared to the instantaneous flow rate per unit time of fuel gas flowing through the large ejector 41 (q1 + q2), further enhancing the drainage effect of MEA stagnant water W1.

[0043] On the other hand, in order to effectively discharge the stagnant water W2 in the flow path to the outside of the fuel cell stack 100, the discharge speed must be set to a predetermined value Va (Figure 8) or higher, and to achieve this, it is necessary to ensure a sufficient acceleration time for the stagnant water W2 in the flow path. In this embodiment, the opening time Ti2 of the small injector 32 is set to be longer than the opening time Ti1 of the large injector 31, so that a sufficient acceleration time can be ensured by the injection of the small injector 32. As a result, the stagnant water W2 in the anode flow path (Figure 3) can be effectively discharged along the flow of fuel gas.

[0044] Figure 7B shows another example of the fuel gas injection pattern from injectors 31 and 32 in the fuel cell system 10 according to this embodiment. In the example of Figure 7B, the small injector 32 is turned on at time tb when the large injector 31 is turned off. In the example of Figure 7B as well, a relatively large instantaneous amount of fuel gas is supplied via the large ejector 41, so the MEA stagnant water W1 can be discharged effectively. Also, the time from when the large injector 31 is turned on until the small injector 32 is turned off is longer than in Figure 7A. As a result, the fuel gas flows through the anode flow path 33 for a longer period of time, and the flow path stagnant water W2 can be discharged effectively.

[0045] In the examples shown in Figures 7A and 7B, the small injector 32 does not turn on after a predetermined time has elapsed since the large injector 31 was turned off, but rather the small injector 32 turns on simultaneously with the turn on or turn off of the large injector 31. In other words, the small injector 32 turns on between the time the large injector 31 is turned on and the time it is turned off. To put it another way, the small injector 32 turns on immediately after the large injector 31 is turned on. This allows for the following additional effects to be achieved.

[0046] Figure 8 shows the change in differential pressure between the fuel gas inlet (through-hole 102a) and outlet (through-hole 102f) over time, and the change in the velocity (drainage velocity V) of the water W2 stagnant in the anode flow path 33. In the figure, the change in differential pressure is shown by a dotted line, and the change in drainage velocity V is shown by a solid line. Figure 8 corresponds to the case in Figure 7B, where the small injector 32 is turned on at the same time as the large injector 31 is turned off. Note that the valve opening time Ti1 of the large injector 31 in Figure 8 is shorter than that in Figure 7A, and the duty cycle of the large injector 31 in Figure 8 is also smaller than that in Figure 7A.

[0047] As shown in Figure 8, when the large injector 31 is turned on at time ta, the differential pressure (dotted line) rises instantaneously, and the drainage velocity V (solid line) also rises accordingly. Subsequently, the differential pressure decreases when the large injector 31 is turned off, but since the small injector 32 is turned on at time tb, the differential pressure does not drop to 0 and is maintained at a predetermined value for the duration of the small injector 32's valve opening time Ti2. Therefore, although the drainage velocity V decreases beyond time tb, the decrease is small and it is maintained at a value at least higher than the predetermined value Va until time tc, when the valve opening time Ti2 ends. In other words, from time ta when the large injector 31 is turned on to time tc when the small injector 32 is turned off, a differential pressure of a predetermined value or higher is continuously generated in the fuel cell stack 100, so the velocity energy of the water flowing through the anode channel 33 is maintained at a predetermined value Va or higher. As a result, there is no need to accelerate the stagnant water from 0 to a predetermined value Va or higher, and efficient drainage of the stagnant water W2 in the channel is possible. The predetermined value Va is the minimum drainage rate required to smoothly discharge the stagnant water W2 in the flow path.

[0048] Figure 9 is a comparative example of Figure 8. In Figure 9, the small injector 32 turns on at time tb, after time Tc has elapsed since the large injector 31 turned off. As a result, the differential pressure (dotted line) becomes 0 and the drainage velocity V (solid line) also becomes 0 before the small injector 32 turns on. Therefore, continuous drainage is interrupted from the time the large injector 31 turns on until the small injector 32 turns off. Then, after the differential pressure starts to rise again at time tb, the drainage velocity V gradually increases. Thus, if the time Tc (referred to as the delay time for convenience) from the time the large injector 31 turns off until the small injector 32 turns on is long, the drainage velocity V decreases significantly (for example, to 0). For this reason, raising the drainage velocity V to a predetermined value Va (Figure 8) or higher requires time and a large amount of energy, making efficient drainage difficult.

[0049] In this regard, in the example shown in Figure 8, the small injector 32 is turned on simultaneously with the large injector 31 being turned off, so the delay time Tc is 0, and a drainage velocity V of 0 or higher than a predetermined value Va can be maintained. Therefore, there is no need to accelerate the stagnant water to 0 or higher than a predetermined value Va, and no energy is required to increase the drainage velocity V, thus enabling efficient drainage.

[0050] Figure 10 shows the relationship between the delay time Tc and the drainage velocity V of the anode flow path 33. As shown in Figure 10, the drainage velocity V gradually decreases as the delay time Tc increases. The delay time Tc at which the drainage velocity V is a predetermined value Va is called the predetermined delay time Tc1. The predetermined delay time Tc1 is stored in the memory of the ECU 50 in advance. The ECU 50 determines the injection timing of the small injector 32 so that it turns on the small injector 32 after the large injector 31 is turned off and before the predetermined delay time Tc1 has elapsed. Note that when the small injector 32 is turned on before the predetermined delay time Tc1 has elapsed, this includes the case where the small injector 32 is turned on at the same time as the large injector 31, as shown in Figure 7A, and the case where the small injector 32 is turned on when the large injector 31 is turned off, as shown in Figure 7B. In other words, it includes the case where there is no delay time Tc.

[0051] The fuel cell system 10 is provided with multiple driver circuits (not shown) to supply drive current to the injectors 31 and 32. However, a predetermined delay time Tc may be required due to constraints on the operation of these driver circuits. In this case, a lower limit of the delay time Tc is set in advance, and the ECU 50 sets the delay time Tc to be greater than or equal to the lower limit and less than or equal to the predetermined delay time Tc1.

[0052] Figure 11 is a flowchart showing an example of a process executed by the ECU 50 according to a program pre-stored in memory. The process shown in this flowchart is the process when fuel gas is injected from injectors 31 and 32 in both injection modes, and the process in Figure 11 is executed repeatedly at predetermined timings.

[0053] As shown in Figure 11, first, in step S1, the required power is calculated based on the signal from the accelerator opening sensor 51. Next, in step S2, the target fuel gas injection amount per unit time for all injectors 31 and 32 is calculated according to the calculated required power. Next, in step S3, the target duty cycle of the large injector 31 and the target duty cycle of the small injector 32 are calculated so that the target injection amount calculated in step S2 is injected. For example, the target duty cycle is calculated along with the target period by referring to pre-stored characteristics.

[0054] Figure 12 shows the relationship between the period and valve opening time in both injection modes that are stored in advance. In the figure, characteristic f1 is the characteristic of the large injector 31, and characteristic f2 is the characteristic of the small injector 32. Both characteristics f1 and f2 are upward sloping, and the valve opening time Ti increases as the period T0 increases. The range from the lower limit period T0a to the upper limit period T0b in the figure is the control range, and the ECU 50 calculates the target period and target duty cycle within this control range. Alternatively, the ECU 50 may calculate the target period and target duty cycle using a predetermined calculation formula.

[0055] Next, in step S4, the ECU 50 outputs a control signal to the large injector 31 (more specifically, the driver circuit) so that the large injector 31 turns on and off at the target cycle and according to the target duty cycle. As a result, the large injector 31 opens, and a large flow rate q1 (Figures 7A and 7B) of fuel gas is instantaneously supplied to the fuel cell stack 100 via the large ejector 41. This allows the MEA stagnant water W1 accumulated near the anode electrode to be effectively guided to the anode flow path 33.

[0056] Next, in step S5, the ECU 50 outputs a control signal to the small injector 32 (more specifically, the driver circuit) so that the small injector 32 turns on and off at the target cycle and according to the target duty cycle. In this case, the control signal is output so that the small injector 32 turns on before a predetermined delay time Tc1 has elapsed after the large injector 31 turns off. For example, the ECU 50 outputs a control signal so that the small injector 32 turns on simultaneously with the turn on of the large injector 31, or so that the small injector 32 turns on simultaneously with the turn off of the large injector 31. In this case, the predetermined delay time Tc1 is 0 in both cases. As a result, the small injector 32 turns on immediately after the turn on of the large injector 31, so that the drainage rate V of the stagnant water W2 in the flow path can be kept at or above a predetermined value Va from the time the large injector 31 turns on until the small injector 32 turns off (Figure 8), and the stagnant water W2 in the flow path in the anode flow path can be discharged well.

[0057] Figure 13 shows the range of power generation conditions that allow for stable power generation. In the figure, the horizontal axis represents the instantaneous flow velocity of the fuel gas, corresponding to the gas flow rates q1 and q2 in Figure 7A. The vertical axis represents the average gas flow rate obtained by dividing the total flow rate of fuel gas supplied to the fuel cell stack 100 over a predetermined time (e.g., period T0) (the sum of the injection amount Q0 and circulation amounts Q11 and Q21 in Figure 6) by the predetermined time (e.g., T0). The area to the right of characteristic f3 is the stable power generation region AR1, and the area to the left is the unstable power generation region AR2.

[0058] According to characteristic f3, the average gas flow rate decreases as the instantaneous gas flow velocity increases. Therefore, if stable power generation is to be achieved using only injection by the large injector 31, the average gas flow rate is small, and the control target value is set within region AR11 in Figure 13, for example. In this case, the range of the control target is narrow, which imposes many constraints on the fuel cell system and leads to increased costs. In contrast, in this embodiment, fuel gas is injected not only using the large injector 31 but also using the small injector 32, which has a high circulation flow rate ratio, so the average gas flow rate increases. As a result, the control target value can be set within region AR12 in Figure 13, for example. Therefore, it is not necessary to increase the instantaneous gas flow path so much, and stable power generation can be easily achieved while keeping cost increases down.

[0059] This embodiment can provide the following effects and advantages. (1) The fuel cell system 10 generates electricity through an electrochemical reaction between a fuel gas and an oxidizer gas and includes a fuel cell stack 100 having a through-hole 102a as a gas inlet for fuel gas to flow in and a through-hole 102f as a gas outlet for fuel gas to flow out; a gas supply unit SP1 having a large injector 31 for injecting fuel gas and supplying fuel gas to the fuel cell stack 100 via a large channel PA11 connected to the through-hole 102a; a gas supply unit SP2 having a small injector 32 for injecting fuel gas and supplying fuel gas to the fuel cell stack 100 via a small channel PA12 connected to the through-hole 102a; and an ECU 50 that controls the large injector 31 and the small injector 32 to inject fuel gas in an injection mode according to the required power (Figure 4). The gas supply unit SP1 is provided between the large injector 31 and the through-hole 102a and further includes a large ejector 41 that guides the fuel gas discharged from the through-hole 102f to the large flow path PA11 via the circulation flow path PA21 (Figure 4). The gas supply unit SP2 is provided between the small injector 32 and the through-hole 102a and further includes a small ejector 42 that guides the fuel gas discharged from the through-hole 102f to the small flow path PA12 via the circulation flow path PA22 (Figure 4). The large injector 31 and the small injector 32 are configured such that the amount of fuel gas injected per unit time q10 by the large injector 31 is greater than the amount of fuel gas injected per unit time q20 by the small injector 32 (Figure 6). The large ejector 41 and the small ejector 42 are configured such that the circulation flow rate ratio (first circulation flow rate ratio), which is the ratio of the circulation amount q11 in the large ejector 41 to the fuel gas injection amount q10 in the large injector 31, is smaller than the circulation flow rate ratio (second circulation flow rate ratio), which is the ratio of the circulation amount q21 in the small ejector 42 to the fuel gas injection amount q20 in the small injector 32 (Figure 6). The ECU 50 controls the large injector 31 and the small injector 32 so that when the large injector 31 injects fuel gas at a predetermined cycle T0, the small injector 32 injects fuel gas for a longer injection time than the large injector 31 during that predetermined cycle T0 (Figures 7A and 7B).

[0060] This configuration allows not only the stagnant water W2 accumulated in the anode channel 33 but also the MEA stagnant water W1 accumulated inside the membrane electrode assembly 20 to be drained from the fuel cell stack 100 at a predetermined cycle T0, thereby optimizing drainage performance. This improves the drainage performance of the fuel cell stack 100 and suppresses power generation instability and electrode performance degradation. Specifically, when the large injector 31 is turned on, the instantaneous gas flow rate increases, allowing for good drainage of the MEA stagnant water W1. However, because the fuel gas injection time is short, sufficient acceleration time for the stagnant water W2 cannot be secured, and the stagnant water W2 cannot be drained effectively. On the other hand, since the opening time Ti2 of the small injector 32 is longer than the opening time Ti1 of the large injector 31 (Figure 12), turning on the small injector 32 improves the drainage performance of the stagnant water W2. However, because the instantaneous gas flow rate is small, the MEA stagnant water W1 cannot be sufficiently drained. In this embodiment, since fuel gas is injected using both injection modes, which combine injection from the large injector 31 and injection from the small injector 32, the MEA stagnant water W1 and the flow path stagnant water W2 can be discharged effectively.

[0061] (2) The ECU 50 controls the large injector 31 and the small injector 32 so that the large injector 31 injects fuel gas at the same time as the large injector 31 injects fuel gas, or within a predetermined time (e.g., valve opening time Ti1) after the large injector 31 injects fuel gas (Figures 7A and 7B). In order to smoothly discharge stagnant water from the fuel cell stack 100, the drainage speed V must be set to a predetermined value Va or higher. In this embodiment, the drainage speed V can be maintained at a predetermined value Va or higher from the start of injection by the large injector 31 until the end of injection by the small injector 32, thereby enabling efficient drainage.

[0062] (3) The ECU 50 controls the large injector 31 and the small injector 32 so that the small injector 32 injects fuel gas before the amount of fuel gas injected by the large injector 31 becomes zero (Figure 8). For example, the small injector 32 is turned on before a predetermined delay time Tc1 has elapsed after the large injector 31 has turned off. This suppresses the decrease in the drainage velocity V and enables efficient drainage.

[0063] (4) The large injector 31 is composed of multiple injectors 31A to 31C, which have the same configuration as the small injector 32 (Figure 4). This reduces the number of parts and allows the fuel cell system 10 to be constructed at a low cost.

[0064] By the way, in the fuel cell system 10 of this embodiment, the amount of fuel gas supplied changes in accordance with the increase or decrease in the required power, and the amount of generated water also changes accordingly. For this reason, it is preferable to improve the wastewater performance by performing different control during transient operation when the required power increases or decreases compared to non-transient operation. Taking this into consideration, in this embodiment, the fuel cell system 10 is further configured as follows.

[0065] Figures 14A and 14B are time charts showing examples of operation during transient operation of the fuel cell system 10 according to this embodiment, respectively. They show the changes over time of the fuel gas flow rate per unit time after passing through the large ejector 41 (large EJE flow rate) Q1, the fuel gas flow rate per unit time after passing through the small ejector 42 (small EJE flow rate) Q2, the amount of generated water Q3 produced in the membrane electrode assembly 20, and the current A flowing to the load (driving motor). Figure 14A shows the operation when the required power increases due to an increase in the amount the accelerator pedal is pressed, for example, and Figure 14B shows the operation when the required power decreases due to a decrease in the amount the accelerator pedal is pressed.

[0066] In Figures 14A and 14B, the large EJE flow rate Q1 is the value obtained by dividing the flow rate of fuel gas flowing through the large ejector 41 when the large injector 31 injects fuel gas once by the period T0, and corresponds to the average flow rate of fuel gas flowing through the large ejector 41. In Figures 14A and 14B, the small EJE flow rate Q2 is the value obtained by dividing the flow rate of fuel gas flowing through the small ejector 42 when the small injector 32 injects fuel gas once by the period T0, and corresponds to the average flow rate of fuel gas flowing through the small ejector 42. Similar characteristics are obtained even if the average injection amount of the large injector 31, obtained by dividing the injection amount of the large injector 31 by the period T0, and the average injection amount of the small injector 32, obtained by dividing the injection amount of the small injector 32 by the period T0, are used instead of the large EJE flow rate Q1 and small EJE flow rate Q2.

[0067] Figures 14A and 14B are time charts showing the process when a predetermined process is executed in the ECU 50. Figures 15A and 15B are flowcharts showing an example of the process executed in the ECU 50 according to a predetermined program, and correspond to the time charts in Figures 14A and 14B, respectively. In order to execute the processes in Figures 15A and 15B, it is necessary to detect the pressure P of the fuel gas at the inlet of the fuel cell stack 100. This pressure P is detected by a pressure sensor 52 connected to the flow path downstream of the ejectors 41 and 42, for example, as shown in Figure 4.

[0068] The flowcharts in Figures 15A and 15B indicate that the process begins when predetermined control start conditions are met. Specifically, the ECU 50 determines, based on the signal from the accelerator pedal position sensor 51, whether the degree of increase in requested power is greater than or equal to a predetermined value. More specifically, it determines whether the requested power has increased by more than or equal to a predetermined value, or whether the rate of increase in requested power is greater than or equal to a predetermined value. When the ECU 50 determines that the degree of increase in requested power is greater than or equal to a predetermined value, it determines that the control start conditions have been met and begins the process shown in Figure 15A.

[0069] As shown in Figure 15A, the ECU 50 first calculates the target pressure Pa of the fuel gas at the inlet of the fuel cell stack 100 in step S11. In this case, in order to increase the inlet pressure P early, the ECU 50 calculates the target pressure Pa to be a value that is a predetermined value or a predetermined percentage higher than the control pressure Pb corresponding to the required power determined by the accelerator pedal opening. The target pressure Pa is larger the greater the increase in required power.

[0070] Next, in step S12, the ECU 50 determines whether the pressure P detected by the pressure sensor 52 is less than the target pressure Pa. If the result in step S12 is positive, the process proceeds to step S13, where the ECU 50 instantaneously increases the duty cycle of the large injector 31 to a target value corresponding to the target pressure Pa, while keeping the duty cycle of the small injector 32 constant. As a result, as shown in Figure 14A, the large EJE flow rate Q1 increases instantaneously while the small EJE flow rate Q2 remains constant (time t1). The increase in the duty cycle continues until the pressure P detected by the pressure sensor 52 reaches the target pressure Pa. This increases the amount of generated water produced in the membrane electrode assembly 20, and also increases the current A flowing to the load.

[0071] If step S12 is rejected, the process proceeds to step S14, where the ECU 50 controls the duty cycle of the large injector 31 to a target value corresponding to the control pressure Pb, while keeping the duty cycle of the small injector 32 constant. As a result, as shown in Figure 14A, the large EJE flow rate Q1 decreases to a value corresponding to the control pressure Pb and is maintained at that value. Consequently, the increase in generated water stops, and the increase in current A also stops (time t2). Note that in step S14, the duty cycle of the small injector 32 may also be changed (for example, decreased).

[0072] As the power demand increases, the duty cycle of the large injector 31 is instantaneously increased while the duty cycle of the small injector 32 remains constant (step S13), thereby increasing the instantaneous gas flow rate through the anode channel 33. This allows for efficient discharge of the MEA stagnant water W1 that has increased due to the increase in power generation. In other words, immediately after increasing the fuel gas supply, the generated water accumulates near the membrane electrode assembly 20 rather than in the anode channel 33. Therefore, by instantaneously increasing the injection amount of the large injector 31, which is suitable for draining the MEA stagnant water W1, the stagnant water in the fuel cell stack can be efficiently discharged.

[0073] When the power requirement decreases, the ECU 50 determines, based on the signal from the accelerator pedal position sensor 51, whether the degree of decrease in the power requirement is greater than or equal to a predetermined value. More specifically, it determines whether the power requirement has decreased by more than or equal to a predetermined value, or whether the rate of decrease in the power requirement is greater than or equal to a predetermined value. If it determines that the degree of decrease in the power requirement is greater than or equal to a predetermined value, it determines that the control start condition has been met and starts the process shown in Figure 15B.

[0074] As shown in Figure 15B, the ECU 50 first calculates the target pressure Pa of the fuel gas at the inlet of the fuel cell stack 100 in step S21. In this case, in order to reduce the inlet pressure P early, the ECU 50 calculates the target pressure Pa to be lower by a predetermined value or a predetermined percentage lower than the control pressure Pb corresponding to the required power determined by the accelerator pedal opening. The target pressure Pa is smaller the greater the decrease in required power.

[0075] Next, in step S22, the ECU 50 determines whether the pressure P detected by the pressure sensor 52 is greater than the target pressure Pa. If the result in step S22 is positive, the process proceeds to step S23, where the ECU 50 instantaneously reduces the duty cycle of the large injector 31 to a target value corresponding to the target pressure Pa, while keeping the duty cycle of the small injector 32 constant. As a result, as shown in Figure 14B, the large EJE flow rate Q1 decreases instantaneously while the small EJE flow rate Q2 remains constant (time t3). The reduction in the duty cycle continues until the pressure P detected by the pressure sensor 52 reaches the target pressure Pa. This reduces the amount of generated water produced in the membrane electrode assembly 20, and also reduces the current A flowing to the load.

[0076] If the step in S22 is rejected, the process proceeds to step S24, where the ECU 50 controls the duty cycle of the large injector 31 to a target value corresponding to the control pressure Pb, while keeping the duty cycle of the small injector 32 constant. As a result, as shown in Figure 14B, the large EJE flow rate Q1 increases to a value corresponding to the control pressure Pb and is maintained at that value. Consequently, the decrease in generated water stops, and the decrease in current A also stops (time t4). Note that in step S24, the duty cycle of the small injector 32 may also be changed (for example, increased).

[0077] As the power demand decreases in this way, the duty cycle of the large injector 31 is instantaneously reduced while the duty cycle of the small injector 32 remains constant (step S23), thus reducing the instantaneous gas flow rate through the anode channel 33. At this time, the amount of MEA stagnant water W1 generated decreases due to the decrease in power generation, but the instantaneous gas injection amount is reduced in response, so the stagnant water can be efficiently discharged. That is, immediately after the decrease in fuel gas supply, the MEA stagnant water W1 decreases, but the channel stagnant water W2 decreases with a delay. Taking this into consideration, the duty cycle of the small injector 32 is maintained to be suitable for draining the channel stagnant water W2, so that the stagnant water in the fuel cell stack can be efficiently discharged.

[0078] In step S13 of Figure 15A, the duty cycle of the large injector 31 is increased while the duty cycle of the small injector 32 is kept constant. A specific example of this will be explained. Figure 16 shows examples of pulse waveforms of injectors 31 and 32 before changing the duty cycle of the large injector 31, and examples of pulse waveforms after changing the duty cycle (first example, second example). In Figure 16, the injection mode is such that the large injector 31 and the small injector 32 are turned on simultaneously, but the small injector 32 may be turned on after the large injector 31 is turned on, and the injection mode is not limited to the example in Figure 16.

[0079] As shown in Figure 16, before the change in the duty cycle of the large injector 31 (before time t1 in Figure 14A), the large injector 31 and the small injector 32 are both turned on with period T0, the pulse width (open time) of the large injector 31 is Ti1, and the pulse width (open time) of the small injector 32 is Ti2.

[0080] In the first example after the change (after time t1 in Figure 14A), the large injector 31 and the small injector 32 are both turned on with the same period T10 as period T0, and the pulse width of the large injector 31 is Ti11, while the pulse width of the small injector 32 is Ti21. The pulse width Ti21 of the small injector 32 is equal to the pulse width Ti2 before the change. On the other hand, the pulse width Till of the large injector 31 is longer than the pulse width Till before the change. As a result, the duty cycle of the large injector 31 increases while the duty cycle of the small injector 32 remains constant.

[0081] In the modified second example, the large injector 31 and the small injector 32 are both turned on with a period T20 that is shorter than period T0, and the pulse width of the large injector 31 is Ti12, while the pulse width of the small injector 32 is Ti22. The pulse width Ti12 of the large injector 31 is equal to the pulse width Ti1 before the modification. On the other hand, the pulse width Ti22 of the small injector 32 is shorter than the pulse width Ti2 because the period has been shortened. As a result, the duty cycle of the small injector 32 remains constant while the duty cycle of the large injector 31 increases.

[0082] In step S23 of Figure 15B, the duty cycle of the large injector 31 is decreased while the duty cycle of the small injector 32 is kept constant, and the pulse waveform in this case can also be changed in the same way as in Figure 16. In this case, in the first example after modification, the pulse width Ti11 of the large injector 31 becomes shorter than the pulse width Ti1 before modification, while maintaining the same period as before modification. Also, in the second example after modification, the period T20 after modification becomes longer than the period T0 before modification, and the pulse width Ti22 of the small injector 32 becomes longer than before modification by that amount. In Figures 15A and 15B, the duty cycle of the large injector 31 is increased or decreased while the duty cycle of the small injector 32 is kept constant, but the duty cycle of the small injector 32 does not have to be constant. In other words, if the ECU 50 controls the large injector 31 and the small injector 32 such that when the power requirement increases, the increase in the duty cycle of the large injector 31 is greater than the increase in the duty cycle of the small injector 32, and when the power requirement decreases, the decrease in the duty cycle of the large injector 31 is greater than the decrease in the duty cycle of the small injector 32, then the duty cycle of the small injector 32 does not need to be constant.

[0083] In addition to the effects described above, this embodiment can also provide the following further effects. (1) When the injection mode is one of both injection modes in which the large injector 31 and the small injector 32 each inject fuel gas, the ECU 50 controls the large injector 31 and the small injector 32 such that when the required power increases, the average value of the injection amount per unit time increases for the large injector 31 more than for the small injector 32, and when the required power decreases, the average value of the injection amount per unit time decreases for the large injector 31 more than for the small injector 32 (Figures 15A and 15B).

[0084] As a result, when the power requirement increases, the injection volume of the large injector 31 increases preferentially over that of the small injector 32. When the power requirement increases, the amount of MEA stagnant water W1 increases, and then the amount of flow channel stagnant water W2 increases. However, by preferentially increasing the injection volume of the large injector 31, the drainage performance of the MEA stagnant water W1 can be improved, and the stagnant water in the fuel cell stack can be efficiently discharged. Also, when the power requirement decreases, the injection volume of the large injector 31 decreases preferentially over that of the small injector 32. When the power requirement decreases, the amount of MEA stagnant water W1 decreases, and then the amount of flow channel stagnant water W2 decreases. However, by preferentially decreasing the injection volume of the large injector 31, the drainage performance of the flow channel stagnant water W2 can be maintained, and the stagnant water in the fuel cell stack can be efficiently discharged.

[0085] (2) When the injection mode is both injection modes, the ECU 50 controls the injectors 31 and 32 so that the large injector 31 injects fuel gas at a duty cycle (Ti1 / T0) (first target duty cycle) corresponding to the required power, and the small injector 32 injects fuel gas at a duty cycle (Ti2 / T0) (second target duty cycle) corresponding to the required power (Figure 16). Furthermore, when the injection mode is both injection modes, if the required power increases, the ECU 50 increases the duty cycle of the large injector 31 to a greater extent than the increase of the small injector 32, while if the required power decreases, the ECU 50 decreases the duty cycle of the large injector 31 to a greater extent than the decrease of the duty cycle of the small injector 32. As a result, by changing the duty cycle, good drainage performance from the fuel cell stack in response to transient operation can be obtained.

[0086] (3) When the injection mode is both injection modes, the ECU 50 increases the duty cycle of the large injector while keeping the duty cycle of the small injector 32 constant when the required power increases, while decreasing the duty cycle of the large injector 31 while keeping the duty cycle of the small injector 32 constant when the required power decreases (Figures 15A, 15B). This makes it possible to obtain good drainage performance from the fuel cell stack in response to transient operation simply by changing the duty cycle of the large injector, and the configuration is simple.

[0087] In the above embodiment, fuel gas is introduced into the fuel cell stack 100 through a through-hole 102a opened in the end unit 102, but the configuration of the gas inlet is not limited to that described above. In the above embodiment, fuel gas is introduced out of the fuel cell stack 100 through a through-hole 102f opened in the end unit 102, but the configuration of the gas outlet is not limited to that. In the above embodiment, the injector 31 is composed of three injectors 31A to 31C with the same configuration as the injector 32, but the configuration of the injector 31 is not limited to that. For example, instead of the injector 31 being composed of multiple injectors 31A to 31C, it may be composed of a single injector with a large injection volume.

[0088] In the above embodiment, the gas supply unit SP1, acting as the first supply unit, supplies fuel gas to the fuel cell stack 100 via a large injector 31 (first injector) and a large ejector 41 (first ejector) located in the large flow path PA11 (first supply flow path). Furthermore, the gas supply unit SP2, acting as the second supply unit, supplies fuel gas to the fuel cell stack 100 via a small injector 32 (second injector) and a small ejector 42 (second ejector) located in the small flow path PA12 (second supply flow path). However, the configuration of the first and second supply units is not limited to those described above. Reed valves may be interposed between the circulation flow path PA21 (first circulation flow path) and the circulation flow path PA22 (second circulation flow path).

[0089] In the above embodiment, when the injection mode is one of both injection modes (predetermined injection mode) in which the large injector 31 and the small injector 32 inject fuel, the ECU 50, acting as a control unit, increases the duty cycle of the large injector 31 (first target duty cycle) while keeping the duty cycle of the small injector 32 (second target duty cycle) constant when the required power increases, and decreases the first target duty cycle while keeping the second target duty cycle constant when the required power decreases. However, the second target duty cycle may also be changed. In this case, when the required power increases, the degree of increase in the first target duty cycle should be greater than the degree of increase in the second target duty cycle, while when the required power decreases, the degree of decrease in the first target duty cycle should be greater than the degree of decrease in the second target duty cycle. In the above embodiment, injectors 31 and 32 were controlled by PWM control. However, the injectors may be controlled by means other than PWM control, as long as the first and second injectors are controlled such that when the required power increases, the average amount injected per unit time increases for the large injector 31 (first injector) more than for the small injector 32 (second injector), and when the required power decreases, the average amount injected per unit time decreases for the first injector more than for the second injector.

[0090] In the above embodiment, an example of applying the fuel cell system 10 to a vehicle was described, but the fuel cell system of the present invention can also be applied to mobile bodies other than vehicles such as aircraft and ships, robots, and various industrial machines.

[0091] The above description is merely an example, and the present invention is not limited by the embodiments and modifications described above, as long as the features of the present invention are not impaired. It is also possible to arbitrarily combine one or more of the above embodiments and modifications, and to combine modifications with each other.

[0092] 31,32 Injectors, 31A,31B,31C Injectors, 41,42 Ejectors, 50 ECU, 100 Fuel Cell Stack, 102a,102f Through-holes, PA11 Large channel, PA12 Small channel, PA21,PA22 Circulation channels, SP1,SP2 Gas supply section

Claims

1. A fuel cell that generates electricity through an electrochemical reaction between a fuel gas and an oxidizer gas, and has a gas inlet into which the fuel gas flows and a gas outlet into which the fuel gas flows out, A first supply unit having a first injector for injecting fuel gas and supplying fuel gas to the fuel cell via a first supply channel connected to the gas inlet, A second supply unit having a second injector for injecting fuel gas, and supplying fuel gas to the fuel cell via a second supply channel connected to the gas inlet, The system includes a control unit that controls the first injector and the second injector to inject fuel gas in an injection mode corresponding to the required power, The first supply unit is provided between the first injector and the gas inlet and further includes a first ejector that guides the fuel gas discharged from the gas outlet to the first supply unit via a first circulation path. The second supply unit is provided between the second injector and the gas inlet and further includes a second ejector that guides the fuel gas discharged from the gas outlet to the second supply unit via a second circulation path. The first injector and the second injector are configured such that the amount of fuel gas injected per unit time by the first injector is greater than the amount of fuel gas injected per unit time by the second injector. The first ejector and the second ejector are configured such that the first circulation flow rate ratio, which is the ratio of the amount of fuel gas circulated by the first ejector to the amount of fuel gas injected by the first injector, is smaller than the second circulation flow rate ratio, which is the ratio of the amount of fuel gas circulated by the second ejector to the amount of fuel gas injected by the second injector. A fuel cell system characterized in that, when the injection mode is a predetermined injection mode in which the first injector and the second injector each inject fuel gas, the control unit controls the first injector and the second injector such that when the required power increases, the average value of the injection amount per unit time increases for the first injector more than for the second injector, and when the required power decreases, the average value of the injection amount per unit time decreases for the first injector more than for the second injector.

2. In the fuel cell system according to claim 1, The control unit controls the first injector and the second injector so that when the injection mode is the predetermined injection mode, the first injector injects fuel gas at a first target duty cycle corresponding to the required power, and the second injector injects fuel gas at a second target duty cycle corresponding to the required power. Furthermore, the control unit is characterized in that, when the injection mode is the predetermined injection mode, if the required power increases, the degree of increase in the first target duty cycle is greater than the degree of increase in the second target duty cycle, and if the required power decreases, the degree of decrease in the first target duty cycle is greater than the degree of decrease in the second target duty cycle.

3. In the fuel cell system according to claim 2, The fuel cell system is characterized in that, when the injection mode is the predetermined injection mode, the control unit increases the first target duty cycle while keeping the second target duty cycle constant when the required power increases, and decreases the first target duty cycle while keeping the second target duty cycle constant when the required power decreases.

4. In the fuel cell system according to any one of claims 1 to 3, The fuel cell system is characterized in that the control unit injects fuel gas at a predetermined cycle and controls the first injector and the second injector such that the fuel gas injection time by the second injector at the predetermined cycle is longer than the fuel gas injection time by the first injector.

Images