JP2009135029A - Fuel cell system, and mobile body - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make reasonable the execution timing of learning of a control value computing function used in computing of a control value in a battery system having an injector. <P>SOLUTION: The fuel cell system equipped with the injector supplying fuel gas to a fuel cell in a fuel gas supply passage communicating to the fuel cell includes a control means computing a control value with the prescribed control value computing function based on the operation state of the fuel cell and controlling the driving of the injector based on the control value; and a learning means learning the control value computing function of the injector. The injector is driven in the prescribed driving cycle and supplies fuel gas to the fuel cell. The learning of the control value computing function is conducted each driving cycle of the injector when the prescribed condition is satisfied. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a fuel cell system provided with a fuel gas supply device that supplies fuel gas to a fuel cell side in a fuel gas supply channel that communicates with the fuel cell, and a movable body provided with the fuel cell system.

  Currently, a fuel cell system including a fuel cell that receives a supply of reaction gas (fuel gas and oxidizing gas) and generates electric power has been proposed and put into practical use. Such a fuel cell system is provided with a fuel gas supply channel for flowing fuel gas from a fuel supply source such as a hydrogen tank to the fuel cell, and the fuel gas supply channel has a desired flow rate on the fuel cell side, for example. And an injector for supplying fuel gas at a pressure.

  The drive of the injector is strictly controlled by a control device of the fuel cell system. The control device calculates a control value such as a fuel injection time using a control value calculation function of a predetermined control calculation rule based on the operating state of the fuel cell, and controls the drive of the injector based on the control value. Yes. For this control, for example, feedback control such as PI control is often used.

  For example, when the control value is calculated by the PI control, for example, a plurality of control value calculation functions such as a map showing the relationship between the secondary target pressure of the injector and the generated current of the fuel cell are used. However, since these control value calculation functions vary depending on, for example, aging of the fuel cell and injectors, and individual differences, it is required to learn and update these control value calculation functions as appropriate. ing. Usually, when learning a control value calculation function, it is necessary to acquire a numerical value indicating the current operation state and learn using it. For example, the operation state at that time is temporarily unstable. When it is in a state, it is not suitable for learning. For this reason, it has been proposed that the control value calculation function is learned when a predetermined condition that is in a steady state is met (see Patent Document 1).

JP 2007-165183 A

  However, normally, the learning is performed for each shortest control cycle in which a control signal is sent from the control device to the injector under a certain condition. For this reason, as long as the conditions are met, learning is performed at a timing shorter than the drive cycle of the injector. If learning is performed at a timing shorter than the driving cycle, some of the learning is not reflected in the next driving, and the learning may be wasted. Also, if the number of times of learning increases unnecessarily, there is a risk that learning with low accuracy may be performed due to calculation errors or the like. As a result, the accuracy of injector control may be reduced.

  The present invention has been made in view of the above points, and in a fuel cell system equipped with a fuel gas supply device such as an injector, execution of learning of a control value calculation function such as a map used for calculating a control value is performed. Its purpose is to optimize timing.

  The present invention for solving the above problems is a fuel cell system comprising a fuel gas supply channel for supplying fuel gas to the fuel cell side in a fuel gas supply channel leading to the fuel cell. A control value is calculated using a predetermined control value calculation function based on the state, and control means for controlling the drive of the fuel gas supply device based on the control value, and the control value calculation function of the fuel gas supply device are learned Learning means, and the fuel gas supply device is driven at a predetermined driving cycle to supply fuel gas to the fuel cell side, and the learning of the control value calculation function is performed by driving the fuel gas supply device. It is performed when a predetermined condition is satisfied for each period.

  The “fuel gas supply device” includes, for example, an internal flow path that communicates the upstream side and the downstream side thereof, and is movably disposed in the internal flow path. An electromagnetically driven injector comprising a valve element whose opening area can be changed and a valve element driving unit that drives the valve element by electromagnetic driving force, or a valve element via a diaphragm by, for example, air pressure or a motor. Examples include a adjustable pressure regulator such as a driven diaphragm regulator.

  The “control value calculation function” includes calculation expressions and relational expressions necessary for calculating the control value. For example, the fuel cell inlet side gas state (secondary gas state of the fuel gas supply device) and the inlet side Relationship between target gas state (secondary target gas state of fuel gas supply device), relationship between fuel cell inlet side gas state (secondary gas state of fuel gas supply device) and power generation current, fuel gas supply device The relational expression showing the relationship between the primary gas state and the secondary gas state, the relationship between the primary gas state of the fuel gas supply device and the power generation current of the fuel cell, the arithmetic expression of the control law, and the like are included.

  Further, “learning the control value calculation function” includes not only updating the control value calculation function itself but also updating the calculated value by recalculation of the control value calculation function.

  In addition, the “control value” includes, for example, an injection amount, an injection time, a duty ratio, a driving frequency, a driving pulse, and the like when the fuel gas supply device is an electromagnetic drive type injector. When the device is a diaphragm type regulator, an applied pressure (for example, fluid pressure or spring pressure) that urges the valve body in the opening direction or the closing direction via the diaphragm is included.

  According to the present invention, since learning of the control value calculation function is performed for each drive cycle of the fuel gas supply device under a certain condition, useless learning is eliminated and the execution timing of learning of the control value calculation function is appropriate. It becomes. Further, this enables learning with higher accuracy and improves control accuracy of the fuel gas supply device.

  In the fuel cell system, the control means communicates a control signal with the fuel gas supply device in a control cycle shorter than the drive cycle of the fuel gas supply device, and learning of the control value calculation function is performed in the drive cycle. The value relating to the fuel gas supply state or the fuel cell operating state at the control signal communication timing of the control cycle that coincides with the switching timing of the fuel cell or at the control signal communication timing of the control cycle closest to the switching timing of the driving cycle is used. May be performed. In such a case, learning of the control value calculation function is performed using values at substantially constant timing within the driving cycle, and thus stable and highly accurate learning is possible.

  The learning of the control value calculation function may be prohibited at least when the duty ratio, which is the ratio of the output time to the drive cycle of the fuel gas supply device, is equal to or greater than a predetermined threshold value. Thereby, it is possible to prevent the control value calculation function from being learned when the fuel gas supply device is in an unstable operation state.

  The learning of the control value calculation function may be performed for each of a plurality of learning regions corresponding to the fuel gas consumption flow rate of the fuel cell. According to this, since learning is performed according to the actual fuel gas consumption of the fuel cell, learning can be stably performed with high accuracy.

  The fuel gas supply device may be an injector that supplies fuel gas to the fuel cell side.

  According to another aspect of the present invention, there is provided a moving body provided with the fuel cell system.

  According to the present invention, since the execution timing of learning of the control value calculation function used when calculating the control value of the fuel gas supply device is optimized, useless learning is eliminated and higher-accuracy learning is possible. Become. As a result, the fuel gas supply device can be controlled with higher accuracy.

  Hereinafter, a fuel cell system 1 according to an embodiment of the present invention will be described with reference to the drawings. In the present embodiment, an example in which the fuel cell system 1 is applied to an on-vehicle power generation system of a fuel cell vehicle (moving body) will be described.

  First, the configuration of the fuel cell system 1 according to the embodiment of the present invention will be described.

  As shown in FIG. 1, the fuel cell system 1 according to the present embodiment includes a fuel cell 10 that generates power by receiving supply of reaction gas (oxidizing gas and fuel gas), and the fuel cell 10 includes air or the like. An oxidizing gas piping system 2 for supplying oxidizing gas, a hydrogen gas piping system 3 for supplying hydrogen gas as a fuel gas to the fuel cell 10, a control device (control means, learning means) 4 for integrated control of the entire system, and the like are provided. Yes.

  The fuel cell 10 has a stack structure in which a required number of unit cells that generate power upon receiving a reaction gas are stacked. The electric power generated by the fuel cell 10 is supplied to a PCU (Power Control Unit) 11. The PCU 11 includes an inverter, a DC-DC converter, and the like that are disposed between the fuel cell 10 and the traction motor 12. Further, the fuel cell 10 is provided with a current sensor 13 for detecting a current during power generation.

  The oxidant gas piping system 2 includes an air supply passage 21 that supplies the oxidant gas (air) humidified by the humidifier 20 to the fuel cell 10, and an air exhaust that guides the oxidant off-gas discharged from the fuel cell 10 to the humidifier 20. A flow path 22 and an exhaust flow path 23 for guiding the oxidizing off gas from the humidifier 21 to the outside are provided. The air supply passage 21 is provided with a compressor 24 that takes in the oxidizing gas in the atmosphere and pumps it to the humidifier 20.

  The hydrogen gas piping system 3 includes a hydrogen tank 30 as a fuel supply source that stores high-pressure (for example, 70 MPa) hydrogen gas, and a fuel gas supply channel for supplying the hydrogen gas in the hydrogen tank 30 to the fuel cell 10. A hydrogen supply channel 31 and a circulation channel 32 for returning the hydrogen off-gas discharged from the fuel cell 10 to the hydrogen supply channel 31 are provided.

  Instead of the hydrogen tank 30, a reformer that generates a hydrogen-rich reformed gas from a hydrocarbon-based fuel, a high-pressure gas tank that stores the reformed gas generated by the reformer in a high-pressure state, and Can also be employed as a fuel supply source. A tank having a hydrogen storage alloy may be employed as a fuel supply source.

  The hydrogen supply flow path 31 is provided with a shutoff valve 33 that shuts off or allows the supply of hydrogen gas from the hydrogen tank 30, a regulator 34 that adjusts the pressure of the hydrogen gas, and an injector 35. A primary pressure sensor 41 and a temperature sensor 42 that detect the pressure and temperature of the hydrogen gas in the hydrogen supply flow path 31 are provided on the upstream side of the injector 35. Further, on the downstream side of the injector 35 and upstream of the junction between the hydrogen supply flow path 31 and the circulation flow path 32, a secondary side pressure sensor 43 that detects the pressure of the hydrogen gas in the hydrogen supply flow path 31. Is provided.

  The regulator 34 is a device that regulates the upstream pressure (primary pressure) to a preset secondary pressure. In the present embodiment, a mechanical pressure reducing valve that reduces the primary pressure is employed as the regulator 34. The mechanical pressure reducing valve has a structure in which a back pressure chamber and a pressure adjusting chamber are formed with a diaphragm therebetween, and the primary pressure is reduced to a predetermined pressure in the pressure adjusting chamber by the back pressure in the back pressure chamber. Thus, a publicly known configuration for the secondary pressure can be employed.

  In the present embodiment, as shown in FIG. 1, the upstream pressure of the injector 35 can be effectively reduced by arranging two regulators 34 on the upstream side of the injector 35. For this reason, the design freedom of the mechanical structure (a valve body, a housing, a flow path, a drive device, etc.) of the injector 35 can be increased. In addition, since the upstream pressure of the injector 35 can be reduced, it is possible to prevent the valve body of the injector 35 from becoming difficult to move due to an increase in the differential pressure between the upstream pressure and the downstream pressure of the injector 35. be able to. Accordingly, it is possible to widen the adjustable pressure width of the downstream pressure of the injector 35 and to suppress a decrease in responsiveness of the injector 35.

  The injector 35 is an electromagnetically driven on-off valve capable of adjusting the gas flow rate and gas pressure by driving the valve body directly with a predetermined driving cycle with an electromagnetic driving force and separating it from the valve seat. The injector 35 includes a valve seat having an injection hole for injecting gaseous fuel such as hydrogen gas, a nozzle body for supplying and guiding the gaseous fuel to the injection hole, and an axial direction (gas flow direction) with respect to the nozzle body. And a valve body that is movably accommodated and opens and closes the injection hole. The valve body of the injector 35 is driven by a solenoid, for example, and the opening area of the injection hole can be switched between two stages or multiple stages by turning on and off the pulsed excitation current supplied to the solenoid. The gas injection time and gas injection timing of the injector 35 are controlled by a control signal output from the control device 4, whereby the flow rate and pressure of hydrogen gas are controlled with high accuracy. The injector 35 directly opens and closes the valve (valve body and valve seat) with an electromagnetic driving force, and has a high responsiveness because its driving cycle can be controlled to a highly responsive region.

  In the present embodiment, as shown in FIG. 1, the injector 35 is disposed on the upstream side of the junction A <b> 1 between the hydrogen supply flow path 31 and the circulation flow path 32. Further, as shown by a broken line in FIG. 1, when a plurality of hydrogen tanks 30 are employed as the fuel supply source, the hydrogen gas supplied from each hydrogen tank 30 joins more than the part (hydrogen gas joining part A2). The injector 35 is arranged on the downstream side.

  A discharge flow path 38 is connected to the circulation flow path 32 via a gas-liquid separator 36 and an exhaust drain valve 37. The gas-liquid separator 36 collects moisture from the hydrogen off gas. The exhaust / drain valve 37 is operated by a command from the control device 4 to discharge moisture collected by the gas-liquid separator 36 and hydrogen off-gas containing impurities in the circulation passage 32 to the outside. . In addition, the circulation channel 32 is provided with a hydrogen pump 39 that pressurizes the hydrogen off gas in the circulation channel 32 and sends it to the hydrogen supply channel 31 side. The hydrogen off-gas discharged through the exhaust / drain valve 37 and the discharge passage 38 is diluted by the diluter 40 and merges with the oxidizing off-gas in the exhaust passage 23.

  The control device 4 detects an operation amount of an acceleration operation member (accelerator or the like) provided in the vehicle, receives control information such as an acceleration request value (for example, a required power generation amount from a load device such as the traction motor 12), Control the operation of various devices in the system. In addition to the traction motor 12, the load device is an auxiliary device (for example, a compressor 24, a hydrogen pump 39, a cooling pump motor, or the like) necessary for operating the fuel cell 10, and various types of vehicles involved in traveling of the vehicle. It is a collective term for power consumption devices including actuators used in devices (transmissions, wheel control devices, steering devices, suspension devices, etc.), occupant space air conditioners (air conditioners), lighting, audio, and the like.

  The control device 4 is configured by a computer system (not shown). Such a computer system includes a CPU, ROM, RAM, HDD, input / output interface, display, and the like, and various control operations are realized by the CPU reading and executing various control programs recorded in the ROM. It is like that.

  Next, a function for controlling the injector 35 of the control device 4 will be described.

  For example, as shown in FIG. 2, the control device 4 determines the amount of hydrogen gas consumed in the fuel cell 10 based on the operating state of the fuel cell 10 (the current value during power generation of the fuel cell 10 detected by the current sensor 13). The amount (hereinafter referred to as “hydrogen consumption”) is calculated (fuel consumption calculation function: B1). In the present embodiment, the hydrogen consumption is calculated and updated every calculation cycle (for example, control cycle) of the control device 4 using a specific calculation formula representing the relationship between the current value of the fuel cell 10 and the hydrogen consumption. To do.

  Further, the control device 4 determines the target pressure value of the hydrogen gas (the fuel cell 10 at the downstream position of the injector 35) based on the operating state of the fuel cell 10 (the current value during power generation of the fuel cell 10 detected by the current sensor 13). Target gas supply pressure) (target pressure value calculation function: B2). In the present embodiment, using a specific map representing a relationship between the current value (FC current) of the fuel cell 10 and the target pressure value as shown in FIG. The target pressure value at the position where the pressure sensor 43 is disposed (the pressure adjustment position where pressure adjustment is required) is calculated and updated.

  Further, the control device 4 deviates between the target pressure value calculated as shown in FIG. 2 and the pressure value (detected pressure value) at the downstream position (pressure adjustment position) of the injector 35 detected by the secondary pressure sensor 43. The feedback correction flow rate is calculated based on (Feedback correction flow rate calculation function: B3). The feedback correction flow rate is a hydrogen gas flow rate (pressure difference reduction correction flow rate) added to the hydrogen consumption in order to reduce the deviation between the target pressure value and the detected pressure value.

In the present embodiment, the feedback correction flow rate is calculated and updated every calculation cycle of the control device 4 using the PI control law. Specifically, the control device 4 multiplies the deviation (e) between the target pressure value and the detected pressure value by a proportional gain (K _P ) to thereby generate a proportional feedback correction flow rate (proportional term: P = K _P × e). And an integral feedback correction flow rate (integral term: I = K _I × ∫ (e) dt) is calculated by multiplying the time integral value of deviation (∫ (e) dt) by the integral gain (K _I ). Then, a feedback correction flow rate including a value obtained by adding these is calculated.

  The control device 4 has a learning function (learning function: B0) for learning, for example, a control value calculation function used when calculating the injector ejection time. The control device 4 performs learning when a predetermined condition is satisfied, for example, every driving cycle of the injector 35. When performing the learning, for example, a detection value such as a current value of the fuel cell 10 at a predetermined timing is used. The learning value is calculated by using the learning value, and then the control value calculation function is updated at a predetermined timing using the learning value. Here, the drive cycle means a stepped (on / off) waveform cycle representing the open / close state of the injection hole of the injector 35. In the present embodiment, the drive period is set to a constant value by the control device 4.

  In the present embodiment, for example, correction (learning) is performed by adding a learning value to the feedback correction flow rate calculation formula as a control value calculation function. The control device 4 uses, for example, the map shown in FIG. 4, that is, a map representing the relationship between the FC current (hydrogen consumption) and the learning value, to determine the secular change of the injector 35 from the FC current detected by the current sensor 13. A learning value of the feedback correction flow rate considering flow rate variations due to individual differences is obtained.

  In the map shown in FIG. 4, the learning value is switched for each fixed hydrogen consumption zone (regions 1 to 6 divided by a broken line in FIG. 4). That is, the learning value is switched for each of a plurality of learning regions corresponding to the hydrogen consumption of the fuel cell 10.

  Further, as shown in FIG. 2, the control device 4 calculates a feedforward corrected flow rate corresponding to the deviation between the previously calculated target pressure value and the currently calculated target pressure value (feedforward corrected flow rate calculation function: B4). The feedforward correction flow rate is a change in the hydrogen gas flow rate due to the change in the target pressure value (correction flow corresponding to the pressure difference). In the present embodiment, the feedforward correction flow rate is calculated and updated every calculation cycle of the control device 4 using a specific calculation formula representing the relationship between the deviation of the target pressure value and the feedforward correction flow rate.

  Further, the control device 4 determines the static flow rate upstream of the injector 35 based on the gas state upstream of the injector 35 (hydrogen gas pressure detected by the primary pressure sensor 41 and hydrogen gas temperature detected by the temperature sensor 42). (Static flow rate calculation function: B5). In the present embodiment, the static flow rate is calculated for each calculation cycle of the control device 4 using a specific calculation formula representing the relationship between the pressure and temperature of the hydrogen gas upstream of the injector 35 and the static flow rate. Update.

  Further, the control device 4 calculates the invalid injection time of the injector 35 based on the gas state upstream of the injector 35 (pressure and temperature of hydrogen gas) and the applied voltage (invalid injection time calculation function: B6). Here, the invalid injection time means the time required from when the injector 35 receives a control signal from the control device 4 until the actual injection is started. In the present embodiment, the invalid injection time is calculated for each calculation cycle of the control device 4 using a specific map representing the relationship between the pressure and temperature of the hydrogen gas upstream of the injector 35, the applied voltage, and the invalid injection time. And update.

  Further, the control device 4 calculates the injection flow rate of the injector 35 by adding the hydrogen consumption amount, the feedback correction flow rate, and the feedforward correction flow rate (injection flow rate calculation function: B7). Then, the control device 4 calculates the basic injection time of the injector 35 by multiplying the value obtained by dividing the injection flow rate of the injector 35 by the static flow rate by the drive cycle of the injector 35, and the basic injection time and the invalid injection time. Are added to calculate the total injection time of the injector 35 (total injection time calculation function: B8).

  Next, drive control of the injector 35 of the control device 4 will be described. FIG. 5 shows a flowchart of such drive control.

  During normal operation of the fuel cell system 1, hydrogen gas is supplied from the hydrogen tank 30 to the fuel electrode of the fuel cell 10 through the hydrogen supply channel 31, and the air that has been subjected to humidification adjustment passes through the air supply channel 21. Then, power is generated by being supplied to the oxidation electrode of the fuel cell 10. At this time, the power (required power) to be drawn from the fuel cell 10 is calculated by the control device 4, and hydrogen gas and air in an amount corresponding to the amount of power generation are supplied into the fuel cell 10. In the present embodiment, the flow rate and pressure of the hydrogen gas supplied to the fuel cell 10 during such normal operation are controlled with high accuracy.

  First, the control device 4 of the fuel cell system 1 detects a current value at the time of power generation of the fuel cell 10 using the current sensor 13 (current detection step: S1). Next, the control device 4 calculates the amount of hydrogen gas (hydrogen consumption) consumed by the fuel cell 10 based on the current value detected by the current sensor 13 (fuel consumption calculation step: S2).

  Next, the control device 4 calculates the target pressure value of the hydrogen gas at the downstream position (pressure adjustment position) of the injector 35 based on the current value detected by the current sensor 13 (target pressure value calculation step: S3). And the control apparatus 4 calculates the feedforward correction | amendment flow volume corresponding to the deviation of the target pressure value calculated last time and the target pressure value calculated this time (feedforward correction | amendment flow rate calculation process: S4).

  Next, the control device 4 detects the pressure value at the downstream position (pressure adjustment position) of the injector 35 using the secondary pressure sensor 43 (pressure value detection step: S5). Then, the control device 4 calculates a feedback correction flow rate based on a deviation between the target pressure value calculated in the target pressure value calculation step S3 and the pressure value (detected pressure value) detected in the pressure value detection step S5 (feedback). Correction flow rate calculation step: S6).

  Next, the control device 4 uses the hydrogen consumption calculated in the fuel consumption flow rate calculation step S2, the feedforward correction flow rate calculated in the feedforward correction flow rate calculation step S4, and the feedback correction flow rate calculated in the feedback correction flow rate calculation step S6. By adding, the injection flow rate of the injector 35 is calculated (injection flow rate calculation step: S7).

  Next, the control device 4 detects the static gas upstream of the injector 35 based on the pressure of the hydrogen gas upstream of the injector 35 detected by the primary pressure sensor 41 and the temperature of the hydrogen gas upstream of the injector 35 detected by the temperature sensor 42. The static flow rate is calculated (static flow rate calculation step: S8). Then, the control device 4 multiplies the value obtained by dividing the injection flow rate of the injector 35 calculated in the injection flow rate calculation step S7 by the static flow rate calculated in the static flow rate calculation step S8 by the drive cycle of the injector 35. The basic injection time of the injector 35 is calculated (basic injection time calculating step: S9).

  Next, the control device 4 determines the injector 35 based on the pressure of the hydrogen gas upstream of the injector 35 detected by the primary pressure sensor 41, the temperature of the hydrogen gas upstream of the injector 35 detected by the temperature sensor 42, and the applied voltage. The invalid injection time is calculated (invalid injection time calculating step: S10). Then, the control device 4 adds the basic injection time of the injector 35 calculated in the basic injection time calculation step S9 and the invalid injection time calculated in the invalid injection time calculation step S10, thereby calculating the total injection time of the injector 35. Calculate (total injection time calculating step: S11).

  Thereafter, the control device 4 controls the gas injection time and the gas injection timing of the injector 35 by outputting a control signal related to the total injection time of the injector 35 calculated in the total injection time calculating step S11. The flow rate and pressure of hydrogen gas supplied to the are adjusted.

  Next, the learning function B0 of the control device 4 will be described.

  For example, as shown in FIG. 6, the drive cycle T of the injector 35 is longer than the control cycle Tc in which the control device 4 transmits a control signal. The control device 4 performs learning when a predetermined condition is satisfied for each drive cycle T of the injector 35. For example, learning of each driving cycle T is executed at the control signal communication timing of the control cycle Tc that coincides with the switching timing of the driving cycle T, or at the control signal communication timing of the control cycle Tc closest to the switching timing of the driving cycle T. In addition, the calculation for performing learning of each driving cycle T shown below is the control cycle Tc closest to the control signal communication timing of the control cycle Tc that coincides with the switching timing of the driving cycle T or the switching timing of the driving cycle T. The detection value at the control signal communication timing is used.

  Hereinafter, the learning operation of the feedback correction flow rate value for calculating the injector injection time will be described with reference to FIG.

  First, the control device 4 detects the generated current value (FC current value) of the fuel cell 10 with the current sensor 13 (step S20), for example, the map shown in FIG. 8, that is, the fuel cell 10 corresponding to the FC current value. It should be learned from the hydrogen consumption of the fuel cell 10 using a map showing the relationship between the hydrogen consumption, the primary pressure (upstream pressure) of the injector 35, and the learning zone set for each predetermined hydrogen consumption range. A learning zone is obtained (step S21).

  In the present embodiment, as shown in FIG. 8, six learning zones are set for each hydrogen consumption amount, in other words, for each FC current, separated by broken lines.

  Next, as shown in FIG. 9, it is determined whether the change value of the FC current detected by the current sensor 13 is equal to or less than a predetermined value (step S22). When the change value of the FC current exceeds the predetermined value (step S22: NO), the process returns to step S20, and when it is equal to or less than the predetermined value (step S22: YES), the current current zone is entered. It is determined whether a predetermined time has elapsed (step S23). In these steps S22 and S23, it is determined whether or not it is in a steady state based on the current change value and the elapsed time after entering the current current zone.

  If the predetermined time has not elapsed since entering the current current zone (step S23: NO), the process returns to step S20. If the predetermined time has elapsed (step S23: YES), FIG. It is determined whether the FC inlet pressure deviation (deviation between the target pressure value and the detected pressure value) obtained in the same manner as in step S6 is equal to or less than a predetermined value (step S24, FIG. 10). In the process of step S24, it is determined whether the learning value set based on the FC inlet pressure deviation in step S27, which will be described later, can be in an appropriate range, that is, whether the learning value is suitable for learning.

  If it is not in a learnable state (step S24: NO), the process returns to step S20. If it is in a learnable state (step S24: YES), the injector injection time obtained in the same manner as in step S6 in FIG. The previous value and the current value of the feedback correction flow value (F / B value) are integrated (step S25).

  Next, it is determined whether or not the cumulative number is equal to or greater than the predetermined number (step S26). If the cumulative number is less than the predetermined number (step S26: NO), the process returns to step S20. (Step S26: YES), the F / B value of the injector injection time integrated in Step S25 is divided by the number of integrations to obtain an average value. For example, the average value of this F / B value is the current value in the learning zone obtained in Step S21. (S27, FIG. 11).

  The control device 4 learns the feedback correction flow rate value for calculating the injector injection time through the above processing. When this learning is performed for all of the learning zones 1 to 6, a map as shown in FIG. 4 is obtained. This learning result is reflected in the feedback correction flow rate calculation function B3 when calculating the injector injection time, and is reflected in the setting of the injection time that is the control value of the injector 35.

  Further, the update timing of the learning value is the control signal communication timing of the control cycle Tc that coincides with the switching timing of the driving cycle T after a predetermined time (750 ms) has elapsed in the same learning zone as shown in FIG. Alternatively, it is performed at the control signal communication timing of the control cycle Tc closest to the switching timing of the drive cycle T.

  According to the above embodiment, the learning of the feedback correction flow rate value for calculating the injector injection time is performed for each drive cycle T of the injector 35 under a certain condition, so there is no useless learning and feedback. The learning execution timing of the corrected flow rate value is optimized. Further, this enables learning with higher accuracy and improves the control accuracy of the injector 35.

  Further, the learning of the feedback correction flow rate value is executed at the control signal communication timing of the control cycle Tc that coincides with the switching timing of the driving cycle T or the control signal communication timing of the control cycle Tc closest to the switching timing of the driving cycle T. Further, learning calculation is performed using the detected value of the timing. For this reason, since learning is performed using detection values at substantially constant timing within the driving cycle T, it is possible to perform stable and highly accurate learning.

In the above embodiment, the learned control value calculation function calculates the feedback correction flow rate value. However, a specific calculation expression representing the relationship between the current value of the fuel cell 10 and the hydrogen consumption amount, A specific map representing the relationship between the current value of the fuel cell 10 and the target pressure value, a specific arithmetic expression representing the relationship between the deviation of the target pressure value and the feedforward correction flow rate, the pressure of the hydrogen gas upstream of the injector 35, and This is a specific arithmetic expression or map that represents the relationship between temperature and static flow rate, a specific map that represents the relationship between the pressure and temperature of the hydrogen gas upstream of the injector 35, the applied voltage, and the invalid injection time. May be. Further, it may be a constant such as the integral gain K _I of the feedback correction proportional gain of the flow rate value calculation formula of the proportional term to calculate the K _P and the integral term.

  In the above embodiment, the learning of the control value calculation function adds the correction to the current calculation formula, but the present invention can also be applied to the update calculation of the integral term I of the PI control.

  The update calculation of the integral term I, which is learning in this case, is performed when a certain condition is satisfied for each drive cycle T of the injector 35 as shown in FIG. For example, whether or not the update calculation of the integral term I is performed in each drive cycle T (update calculation execution determination) is determined by the control signal communication timing of the control cycle Tc coincident with the switching timing of the drive cycle T or the drive cycle T This is performed at the control signal communication timing of the control cycle Tc closest to the switching timing. Further, the value used for the calculation of the integral term I is the time when the control signal of the control cycle Tc is communicated and the timing coincides with the switching timing of the driving cycle T or the timing closest to the switching timing of the driving cycle T Is used.

  Next, a condition for permitting the update calculation of the integral term I of the feedback correction flow rate related to the injector control will be described with reference to FIG.

  The figure shows a timing chart of FC current value, gas injection command time, injector secondary pressure command value, and injector drive cycle. Each time t3-t1 has shown the injector injection timing. The FC current values I3 to I1 are current values detected by the current sensor 13 at each injector injection timing. The gas injection command times τ3 to τ1 indicate the time during which the fuel gas is injected from the injector 35 at each injector injection timing. The injector secondary pressure command values lo_ref3 to lo_ref1 are target values of the injector secondary pressure at the respective injector injection timings. The injector driving cycle T indicates the gas injection interval of the injector 34. Duty3 to Duty1 are duty ratios that are ratios of output time to drive periods in each drive period T. The injector drive cycle T3 indicates the time interval between time t3 and time t2, and the gas injection command time τ3 indicates the time during which gas is injected during the injector drive cycle T3. Similarly, the injector drive cycle T2 indicates the time interval between time t2 and time t1, and the gas injection command time τ2 indicates the time during which gas is injected during the injector drive cycle T2.

In the present embodiment, the update calculation of the integral term I by the feedback correction flow rate calculation function B3 is permitted on condition that all of the following conditions (1) to (4) are satisfied.
(1) The injector 35 stably injects gas.
(2) The time change amount of the injector secondary pressure command value is less than a predetermined threshold value.
(3) The amount of time change of the FC current is less than a predetermined threshold value.
(4) The duty ratio, which is the ratio of the output time (time when the injector 35 is ON) To with respect to the drive cycle T of the injector 35, is less than the threshold value.
On the other hand, when any of the above conditions (1) to (4) is not satisfied, the integral term update calculation by the feedback correction flow rate calculation function B3 is prohibited.

Here, in order for the condition (1) to be satisfied, it is necessary that each injector injection time is not zero, that is, the expression (1A) must be satisfied.
τ1> 0 and τ2> 0 and τ3> 0 (1A)
When the expression (1A) is established, the injector injection stability flag is turned on. On the other hand, when any one of τ1, τ2, and τ3 is zero, that is, when the formula (1A) is not satisfied, the injector injection stability flag is turned off.

In order for the condition (2) to be satisfied, the time change amount of the injector secondary pressure command value must be less than a predetermined threshold, that is, all of the following expressions (2A) and (2B) must be satisfied. .
Δlo_ref3 = | lo_ref3-lo_ref2 | / T3 ≦ 20 Pa / s (2A)
Δlo_ref2 = | lo_ref2-lo_ref1 | / T2 ≦ 20 Pa / s (2B)
When all of the equations (2A) to (2B) are satisfied, the injector secondary pressure stabilization flag is turned on. On the other hand, if any one of the equations (2A) and (2B) is not satisfied, the injector secondary pressure stabilization flag is turned off.

In order to satisfy the condition (3), it is necessary that the amount of time change of the FC current is less than a predetermined threshold value, that is, all of the following expressions (3A) and (3B) must be satisfied.
ΔI3 = | I3-I2 | / T3 ≦ 30 mA / s (3A)
ΔI2 = | I2−I1 | / T2 ≦ 30 mA / s (3B)
When all of the expressions (3A) to (3B) are established, the FC current stabilization flag is turned on.
On the other hand, if any one of the equations (3A) and (3B) is not established, the FC current stabilization flag is turned off.

In order to satisfy the condition (4), it is necessary that the duty ratio of the injector 35 is less than the threshold value, that is, all of the following expressions (4A) to (4C) must be satisfied.
Duty1 ≦ 75% (4A)
Duty2 ≦ 75% (4B)
Duty3 ≦ 75% (4C)
When all of the equations (4A) to (4C) are satisfied, the command pressure followable flag is turned on. On the other hand, if any one of the equations (4A) to (4C) is not established, the command pressure follow-up flag is turned off.

  When all of the injector injection stability flag, the injector secondary pressure stabilization flag, the FC current stabilization flag, and the command pressure followable flag are turned on, the integration permission flag is turned on, and the integral term is updated by the feedback correction flow rate calculation function B3. Arithmetic is allowed. On the other hand, when any of the injector injection stability flag, the injector secondary pressure stability flag, the FC current stability flag, and the command pressure followable flag is turned off, the integration permission flag is turned off, and the integration by the feedback correction flow rate calculation function B3 is performed. Term update operations are prohibited.

  As shown in FIG. 12, the integral term I update calculation execution determination is performed for each drive cycle T, and when permitted, the integral term I update calculation is performed using the detected value of the switching timing of the drive cycle T. Thus, the value of the new integral term I is reflected in the driving of the next driving cycle T. In the case of prohibition, the value of the integral term I of the previous driving cycle is maintained.

  According to the above embodiment, the update calculation (learning) of the integral term I of the feedback correction flow rate for calculating the injector injection time is performed for each drive cycle T of the injector 35 under a certain condition. Useless learning is eliminated, and the learning execution timing of the integral term I is optimized. Further, this enables learning with higher accuracy and improves the control accuracy of the injector 35.

  The learning of the integral term I uses the detection value at the control signal communication timing of the control cycle Tc that coincides with the switching timing of the drive cycle T or the control signal communication timing of the control cycle Tc closest to the switching timing of the drive cycle T. Has been done. For this reason, since learning is performed using detection values at substantially constant timing within the driving cycle T, it is possible to perform stable and highly accurate learning.

  Further, the update calculation of the integral term I is prohibited when at least the duty ratio, which is the ratio of the output time to the drive cycle T of the injector 35, is greater than or equal to a predetermined threshold value. Since the integral term I is updated when the operation is stable and the operation is stable, the control error can be reduced.

  The preferred embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to such examples. It will be apparent to those skilled in the art that various changes and modifications can be made within the scope of the ideas described in the claims, and these are naturally within the technical scope of the present invention. It is understood that it belongs.

  For example, in the above embodiment, the calculated control value is the ejection time of the injector 35, but other control values such as an injection amount, a duty ratio, a drive frequency, and a drive pulse may be used. In addition, the fuel gas supply device to be controlled is the injector 35, but another fuel gas supply device such as a regulator 34 may be used.

  Further, although the fuel cell system mounted on the fuel cell vehicle has been described in the above embodiment, the fuel cell system is mounted on various mobile bodies (robots, ships, aircrafts, etc.) other than the fuel cell vehicle. May be. Further, the fuel cell system may be applied to a stationary power generation system used as a power generation facility for a building (house, building, etc.).

It is explanatory drawing which shows the outline of a structure of a fuel cell system. It is a functional block diagram of injector control. It is a map which shows the relationship between FC electric current and target pressure. It is a map which shows the relationship between FC electric current (hydrogen consumption) and a learning value. It is a flowchart for demonstrating the operating method of a fuel cell system. It is a timing chart which shows learning execution timing. It is a flowchart for demonstrating the calculation process of the injector injection time in a fuel cell system. It is a map which shows the learning zone corresponding to hydrogen consumption. It is a map of FC electric current and elapsed time which show the state which can be learned. It is a graph which shows the fluctuation | variation of FC input pressure. It is a graph which shows the fluctuation | variation of the F / B value calculated at the time of learning. It is a timing chart which shows the update calculation execution determination timing of an integral term. 5 is a timing chart of an FC current value, a gas injection command time, an injector secondary pressure command value, an injector drive cycle, and an injector duty ratio.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Fuel cell system 4 Control apparatus 10 Fuel cell 31 Hydrogen supply flow path 35 Injector

Claims (6)

A fuel cell system comprising a fuel gas supply channel for supplying fuel gas to a fuel cell side in a fuel gas supply channel leading to the fuel cell,
Control means for calculating a control value using a predetermined control value calculation function based on the operating state of the fuel cell, and for controlling the driving of the fuel gas supply device based on the control value;
Learning means for learning the control value calculation function of the fuel gas supply device,
The fuel gas supply device is driven at a predetermined driving cycle to supply fuel gas to the fuel cell side,
The control value calculation function is learned when a predetermined condition is satisfied for each driving cycle of the fuel gas supply device. The control means communicates a control signal with the fuel gas supply device at a control cycle shorter than the drive cycle of the fuel gas supply device,
The learning of the control value calculation function is performed at a control signal communication timing of the control cycle that coincides with a switching timing of the driving cycle or a control signal communication timing of the control cycle that is closest to the switching timing of the driving cycle. The fuel cell system according to claim 1, wherein the fuel cell system is performed using a value related to a supply state or an operating state of the fuel cell.   The fuel cell system according to claim 1 or 2, wherein learning of the control value calculation function is performed for each of a plurality of learning regions corresponding to a fuel gas consumption flow rate of the fuel cell. The learning of the control value calculation function is an update operation of an integral term in PI control,
The update calculation of the integral term is prohibited when at least a duty ratio, which is a ratio of an output time with respect to a driving cycle of the fuel gas supply device, is a predetermined threshold value or more. Fuel cell system.   The fuel cell system according to claim 1, wherein the fuel gas supply device is an injector that supplies fuel gas to a fuel cell side.   A moving body comprising the fuel cell system according to claim 1.

Images