JP2007012549A - Fuel cell system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To convert a demanded power to an extracted current amount from a fuel cell precisely with a simple structure. <P>SOLUTION: A controller 26 calculates a necessary extracted current amount from a fuel cell in order to output a demanded power in accordance with output characteristics of a standard of a fuel cell stack 1 and learn a relationship between output characteristics of a standard of the fuel cell stack 1 and actual output characteristics of the fuel cell stack 1. Thus the controller 26 calculates a corrected amount of an extracted current based on the learning results and correct an extracted current amount based on the corrected amount, and control a flowing volume and pressure of hydrogen and air to be supplied to the fuel cell stack 1 by making the corrected extracted current amount as a target current value. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a fuel cell system having a fuel cell that generates power upon receiving a reaction gas.

In general, a system for calculating the amount of extraction current from a fuel cell necessary for supplying the required power and controlling the amount of reaction gas supplied to the fuel cell according to the calculated amount of extraction current is an output that serves as a reference for the fuel cell. Based on the characteristics, an operation is performed for converting the required power into an extracted current amount. However, since the output characteristics of fuel cells usually change due to individual differences in fuel cells and changes over time, when calculating the amount of extraction current based on the output characteristics that serve as the reference for fuel cells, the calculation accuracy is fuel. It changes due to individual differences of batteries and changes over time. Against this background, recently, an approximate expression indicating the output characteristics of the fuel cell has been derived, and the coefficients of each term of the approximate expression derived using a general learning algorithm such as the least squares method are repeatedly used as learning parameters. A system that learns the output characteristics of a fuel cell by calculation has been proposed (see Patent Document 1).
JP 2000-357526 A

  However, converting the output current of the fuel cell to output power based on the output characteristics is easy to calculate because it is only necessary to multiply the output current and the output voltage at that time. In order to do so, it is necessary to perform a convergence operation such as calculating a solution of a higher-order equation or searching for an extraction current value that is a required power value, and thus high calculation processing capability is required. For this reason, in a system that performs an operation using an information processing device having a low calculation capability such as a microcomputer, the required power cannot be extracted on-board and converted into an extraction current amount in real time. For this reason, according to the conventional system, it is difficult to appropriately control the amount of reaction gas supplied to the fuel cell according to the amount of extraction current and to output the required power with high accuracy.

  In order to solve such a problem, a method of learning the characteristics of the extraction current amount of the fuel cell with respect to the required power is conceivable, but even when this method is used, an approximate expression indicating the characteristics is high. Next, it is necessary to perform learning in consideration of changes in the operating temperature, so that the learning algorithm becomes complicated, and it is difficult to take the required power on-board and convert it into the extracted current amount in real time.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a fuel cell system capable of accurately converting required power into an amount of current taken from a fuel cell with a simple configuration. There is to do.

  In order to solve the above-described problem, the fuel cell system according to the present invention calculates the amount of current taken from the fuel cell that is required to output the required generated power based on the output characteristics that serve as a reference for the fuel cell, The relationship between the fuel cell reference output characteristic and the actual output characteristic of the fuel cell is learned, and the extraction current amount is corrected based on the learning result.

  According to the fuel cell system of the present invention, it is not necessary to perform a complicated calculation process such as a convergence calculation in order to convert the required power into the amount of current taken from the fuel cell. Can be accurately converted into the amount of current extracted from the.

  Hereinafter, the configuration and operation of the fuel cell system according to the first and second embodiments of the present invention will be described with reference to the drawings.

[Configuration of fuel cell system]
The fuel cell system according to the first embodiment of the present invention is used as a driving power source of a vehicle such as a hybrid electric vehicle, and as shown in FIG. 1, as a fuel gas and an oxidant gas at an anode and a cathode, respectively. A fuel cell stack 1 is provided in which a plurality of fuel cells that generate power by receiving supply of hydrogen and air are stacked. In this embodiment, the fuel cell includes an electrolyte membrane sandwiched between an anode and a cathode, and the electrolyte membrane is made of a solid polymer electrolyte membrane in consideration of high energy density, low cost, and light weight. Is formed. The solid polymer electrolyte membrane is made of a polymer membrane having ion (proton) conductivity, such as a fluororesin ion exchange membrane, and functions as an ion conductive electrolyte when saturated with water. The electrochemical reaction at the anode and the cathode and the electrochemical reaction as the fuel cell stack 1 as a whole are based on the following formulas (1) to (3).

[Anode] H ₂ → 2H ⁺ + 2e ⁻ (1)
[Cathode] 1/2 O ₂ + 2H ⁺ + 2e ⁻ → H ₂ O (2)
[Overall] H ₂ +1/2 O ₂ → H ₂ O (3)
[Configuration of hydrogen system]
The fuel cell system includes a high-pressure hydrogen tank 2 and a hydrogen pressure control valve 3, and the hydrogen pressure supplied from the high-pressure hydrogen tank 2 is reduced to a pressure suitable for the operating state of the fuel cell stack 1 by the hydrogen pressure control valve 3. After that, hydrogen is supplied to the fuel cell stack 1 through the hydrogen supply channel 4. The hydrogen pressure supplied to the fuel cell stack 1 is controlled by driving the hydrogen pressure control valve 3 by feeding back the hydrogen pressure detected by the pressure sensor 19.

  Unused hydrogen in the fuel cell stack 1 is circulated to the hydrogen inlet side of the fuel cell stack 1 through the hydrogen circulation channel 5 and the ejector 6. By providing the hydrogen circulation channel 5 and the ejector 6, it becomes possible to reuse unused hydrogen in the fuel cell stack 1 and improve the fuel efficiency of the fuel cell system. In the hydrogen circulation channel that returns to the fuel cell stack 1 via the hydrogen circulation channel 5 and the ejector 6, impurity gases such as nitrogen and argon in the air leaked from the cathode side, or excessive moisture was liquefied. Liquid water may accumulate. These impurity gases lower the partial pressure of hydrogen to reduce power generation efficiency, or increase the average molecular weight of the circulating gas, making it difficult to circulate hydrogen. Liquid water also hinders hydrogen circulation.

  For this reason, a hydrogen discharge flow path 7 and a hydrogen purge valve 8 for opening and closing the hydrogen discharge flow path 7 are provided on the hydrogen outlet side of the fuel cell stack 1. When the impurity gas or liquid water accumulates, the hydrogen purge valve 8 is opened for a short time to perform a purge for discharging the impurity gas or liquid water out of the system. As a result, the hydrogen partial pressure and the circulation performance in the hydrogen circulation passage including the fuel cell stack 1 can be recovered.

[Air system configuration]
The fuel cell system includes a compressor 10 that pumps air to the fuel cell stack 1 via the air supply passage 9 and a system that adjusts the pressure of the air discharged from the fuel cell stack 1 via the air discharge passage 11. And an air pressure control valve 12 for discharging to the outside. The pressure of the air supplied to the fuel cell stack 1 is controlled by driving the air pressure control valve 12 by feeding back the air pressure detected by the pressure sensor 21.

[Cooling system configuration]
The fuel cell system includes a coolant circulation pump 14 that pumps the coolant to the fuel cell stack 1 via the coolant flow path 13, and a coolant circulation pump 14 that cools the coolant discharged from the fuel cell stack 1. And a heat exchanger 15 for supplying to the heat exchanger.

[Control system configuration]
The fuel cell system includes a temperature sensor 16 and a pressure sensor 17 that detect the temperature and pressure of hydrogen in the high-pressure hydrogen tank 2, and a temperature sensor 18 and a pressure that detect the temperature and pressure of hydrogen supplied to the fuel cell stack 1. A sensor 19, a flow sensor 20 and a pressure sensor 21 for detecting the flow rate and pressure of air supplied to the fuel cell stack 1, a temperature sensor 22 for detecting the temperature of the coolant discharged from the fuel cell stack 1, and a fuel A voltage sensor 23 for detecting the output voltage of the battery stack 1, a current sensor 24 for detecting the output current of the fuel cell stack 1, a power control device 25 for controlling the amount of power taken out from the fuel cell stack 1, and the entire fuel cell system And a controller 26 for controlling the operation of the above.

  In this embodiment, the power control device 25 is constituted by a DC / DC converter. The DC / DC converter has different switching elements that operate at the time of step-up conversion and step-down conversion, and can output a desired voltage by changing the duty ratio of a control signal applied to the switching element. Specifically, the switching element is controlled to output a voltage equal to or higher than the input voltage at the time of step-up, and the switching element is controlled to output a voltage equal to or lower than the input voltage at the time of step-down.

  The controller 26 is constituted by a microprocessor having a CPU, a program ROM, a working RAM, and an input / output interface, and the CPU executes a control program stored in the program ROM. 4 functions as the required generated power calculation means 31, the actual current detection means 32, the actual voltage detection means 33, the output characteristic learning means 34, the extraction current correction amount calculation means 35, and the extraction current calculation means 36 according to the present invention. The functions of these units will be described later.

  Further, the operating pressure of the fuel cell stack 1 is a variable pressure, and the operating pressure is high when the output taken out from the fuel cell stack 1 is high, and conversely, the operating pressure is low when the output is low. In the fuel cell system having such a configuration, the controller 26 executes the power control process described below, thereby converting the required power into the amount of current taken out from the fuel cell stack 1 with a simple configuration. The operation of the controller 26 when executing this power control process will be described below with reference to the flowchart shown in FIG.

[Power control processing]
The flowchart shown in FIG. 3 starts in response to the start of the fuel cell system, and the power control process proceeds to step S1. This power control process is repeatedly executed every predetermined control cycle (for example, 10 [msec] cycle) after activation.

  In the process of step S1, the controller 26 functions as the required generated power calculation means 31, thereby calculating the generated power required from the electric load connected to the fuel cell system (hereinafter referred to as required generated power). The details of the required generated power calculation process will be described later with reference to the flowchart shown in FIG. Thereby, the process of step S1 is completed and a power control process progresses to the process of step S2.

  In the process of step S2, the controller 26 functions as the output characteristic learning means 34, so that the fuel cell stack 1 is controlled based on the output voltage and output current of the fuel cell stack 1 detected by the voltage sensor 23 and the current sensor 24. Learn current-voltage characteristics. The details of the current-voltage characteristic learning process will be described later with reference to the flowchart shown in FIG. Thereby, the process of step S2 is completed, and the power control process proceeds to the process of step S3.

  In the process of step S3, it is necessary for the controller 26 to function as the extraction current calculation means 36 and to output the required generated power calculated by the process of step S1 based on the current-voltage characteristics that serve as the reference for the fuel cell. The amount of current taken out from the fuel cell stack 1 is calculated. Further, the controller 26 functions as the extraction current correction amount calculation means 35, thereby calculating the correction amount of the extraction current amount based on the learning result of the process of step S2.

  Then, the controller 26 functions as the extraction current calculation means 36 to calculate the target current value of the fuel cell stack 1 based on the calculated extraction current amount and its correction amount. The details of this target current calculation process will be described later with reference to the flowchart shown in FIG. Thereby, the process of step S3 is completed and the power control process proceeds to the process of step S4.

  In the process of step S4, the controller 26 controls the flow rate and pressure of hydrogen and air supplied to the fuel cell stack 1 so as to output the target current value calculated by the process of step S3. The details of the gas supply control process will be described later with reference to the flowchart shown in FIG. Thereby, the process of step S4 is completed and the power control process proceeds to the process of step S5.

  In the process of step S5, the controller 26 controls the power control device 25 so as to take out the required generated power from the fuel cell stack 1. Thereby, the process of step S5 is completed and a series of power control processes are complete | finished.

[Required power generation calculation processing]
The flowchart shown in FIG. 4 starts in response to the start of the fuel cell system, and the required generated power calculation process proceeds to step S11. This required generated power calculation process is repeatedly executed every predetermined control cycle (for example, 10 [msec] cycle) after startup.

  In step S11, the controller 26 detects the accelerator operation amount of the driver based on the output of the accelerator sensor provided in the vehicle. Thereby, the process of step S11 is completed and the required generated power calculation process proceeds to the process of step S12.

  In step S12, the controller 26 detects the speed of the vehicle based on the output of a vehicle speed sensor provided in the vehicle. Thereby, the process of step S12 is completed and the required generated power calculation process proceeds to the process of step S13.

  In the process of step S13, the controller 26 detects the accelerator operation amount and the vehicle detected by the processes of steps S11 and S12 from the map data indicating the relationship between the accelerator operation amount and the required generated power for each vehicle speed as shown in FIG. The required generated power is calculated by reading out the required generated power corresponding to the speed. Thereby, the process of step S13 is completed and a series of request | required power generation calculation processes are complete | finished.

[Current-voltage characteristics learning process]
The flowchart shown in FIG. 6 starts when the process of step S1 is completed, and the current-voltage characteristic learning process proceeds to the process of step S21.

  In the process of step S21, the controller 26 functions as the actual current detection unit 32 and the actual voltage detection unit 33, so that the output voltage and output current (hereinafter referred to as the output voltage and output current) of the fuel cell stack 1 are obtained using the voltage sensor 23 and the current sensor 24. , Expressed as actual voltage and actual current). Thereby, the process of step S21 is completed and the current-voltage characteristic learning process proceeds to the process of step S22.

  In the process of step S <b> 22, the controller 26 detects the temperature of the fuel cell stack 1 based on the coolant temperature detected by the temperature sensor 22. The controller 26 calculates the average value of the coolant temperature at the inlet side and the outlet side of the fuel cell stack 1, and calculates the temperature of the fuel cell stack 1 from the temperature at the location where the fuel cell stack 1 and the operating load are correlated. May be estimated. Thereby, the process of step S22 is completed, and the current-voltage characteristic learning process proceeds to the process of step S23.

  In the process of step S23, the controller 26 determines whether or not the fuel cell stack 1 is in a state where the learning process of the current voltage characteristics can be performed in order to prevent erroneous learning of the current voltage characteristics. If the learning process is not ready to be executed, the controller 26 sets the value of the learning execution permission flag to 0. Conversely, if the learning process is ready to be executed, the controller 26 sets the value of the learning execution permission flag to 1. Set. Details of the learning execution determination process will be described later with reference to the flowchart shown in FIG. Thereby, the process of step S23 is completed, and the current-voltage characteristic learning process proceeds to the process of step S24.

  In the process of step S24, the controller 26 determines whether or not the value of the learning execution permission flag is 1. As a result of the determination, if the value of the learning execution permission flag is 0, the controller 26 determines that the fuel cell stack 1 is not in a state where the learning process of the current voltage characteristics can be performed, and a series of current voltage characteristics learning is performed. The process ends. On the other hand, when the value of the learning execution permission flag is 1, the controller 26 determines that the fuel cell stack 1 is in a state where the learning process of the current voltage characteristics can be performed, and performs the current voltage characteristics learning process in step S25. Proceed to processing.

  In the process of step S25, the controller 26 is detected by the process of steps S21 and S22 from the map showing the relationship between the actual current of the fuel cell stack 1 and the reference voltage for each temperature of the fuel cell stack 1 as shown in FIG. The reference voltage of the fuel cell stack 1 is calculated by reading the reference voltage corresponding to the actual current and temperature. Note that the controller 26 calculates the reference voltage of the fuel cell stack 1 using map data or a function that receives and outputs parameters and a reference voltage that change according to the operating state of the fuel cell stack 1 such as operating pressure. May be. Thereby, the process of step S25 is completed, and the current-voltage characteristic learning process proceeds to the process of step S26.

In the process of step S26, the controller 26 learns the voltage change of the actual current-voltage characteristic with respect to the current-voltage characteristic that is the reference of the fuel cell stack 1 using the learning formula. Specifically, the difference ΔPMIV between the reference voltage and the actual voltage of the fuel cell stack 1 (the hatched area shown in FIG. 8) is expressed by the following formula 1 using the actual current PMIC of the fuel cell stack 1. Therefore, the controller 26 repeatedly calculates the formula 1 as a learning formula and the parameters A (k) and B (k) of the formula 1 as the learning parameters, so that the actual current with respect to the current-voltage characteristic serving as the reference of the fuel cell stack 1 is calculated. Learn voltage changes in voltage characteristics. Thereby, the process of step S26 is completed and a series of current-voltage characteristic learning processes are completed.

  The learning parameters A (k) and B (k) have characteristics that change from moment to moment with respect to changes in current-voltage characteristics of the fuel cell stack 1 over time. The actual voltage of the fuel cell stack 1 has a measurement error due to the influence of the resolution of the voltage sensor 23. Therefore, in this embodiment, the controller 26 uses the RLS (Sequential Least Squares method, see the equation in FIG. 9), which is generally known as an adaptive parameter estimation algorithm used in the field of adaptive control. I do.

  According to RLS, learning parameters A (k) and B (k) can be calculated statistically, so that erroneous learning with respect to measurement errors can be prevented, and forgetting counts in mathematical formulas used in RLS. By setting the value of λ to 0 or more and 1 or less, the weight of the past data with respect to the current data can be changed, so that the learning parameter A (k) with respect to the change with time of the current-voltage characteristics of the fuel cell stack 1 , B (k) can be made to follow.

  Note that the learning parameters A (k) and B (k) may be calculated using other calculation methods such as a maximum likelihood method and a collective least square method. In addition, when performing fixed-point arithmetic, a large arithmetic error occurs during time update of P (k) (see equation (10) in FIG. 9) used in RLS, and as a result, erroneous learning may occur. This is because P (k) must always have positive definiteness and symmetry properties, but when iterative computation is performed at a fixed point, the positive definiteness of P (k) and This is because the symmetry is broken.

  From this, in this embodiment, when performing a fixed-point operation, P (k) is decomposed by UD, and the upper triangular row matrix U and the diagonal component D are respectively updated in time, so that the calculation is performed according to the quantization error. Avoid errors. Through the above processing, when the true values of the learning parameters A (k) and B (k) are PARAM1 and PARAM2, respectively, the learning parameters A (k) and B (k) are set to PARAM1, as shown in FIG. Asymptotically approaches PARAM2.

[Learning execution determination process]
The flowchart shown in FIG. 11 starts when the process of step S22 is completed, and the learning execution determination process proceeds to the process of step S31.

  In the process of step S31, the controller 26 sets the value of the learning execution permission flag to 0, and initializes the learning execution permission flag. Thereby, the process of step S31 is completed and the learning execution determination process proceeds to the process of step S32.

  In the process of step S32, the controller 26 determines whether or not the dispersion value of the actual voltage of the fuel cell stack 1 is equal to or less than a predetermined value. The predetermined value is set to the maximum value of the dispersion value generated due to the influence of the resolution of the voltage sensor 23. However, other values resulting from the operating state of the fuel cell stack 1 such as an actual current, an actual voltage, an operating pressure, a hydrogen utilization rate, and an air utilization rate may be used as the dispersion value. Thereby, the process of step S32 is completed, and the learning execution determination process proceeds to the process of step S33.

  In the process of step S33, the controller 26 determines whether or not the fuel cell stack 1 is in a steady power generation state. Specifically, when the variance value of the actual voltage of the fuel cell stack 1 is less than or equal to a predetermined value, the controller 26 determines that the fuel cell stack 1 is in a steady power generation state, and conversely, the variance value is not less than or equal to the predetermined value. The controller 26 determines that the fuel cell stack 1 is not in the steady power generation state. As a result of the determination, if the fuel cell stack 1 is not in the steady power generation state, the controller 26 determines that the fuel cell stack 1 is not in a state where the current voltage characteristics learning process can be performed, and a series of learning execution determination processes Exit. On the other hand, when the fuel cell stack 1 is in the steady power generation state, the controller 26 advances the learning execution determination process to the process of step S34.

  In the process of step S34, the controller 26 determines whether or not the actual current of the fuel cell stack 1 is a learnable current value. Specifically, as shown in FIG. 12, an extremely low load region in which an activation overvoltage not more than the first predetermined value A appears significantly, or a concentration not less than the second predetermined value B greater than the first predetermined value A. In the extremely high load region in which the overvoltage appears remarkably, the voltage change is large with respect to the current change amount, and the voltage cannot be detected stably.

  Therefore, when the actual current of the fuel cell stack 1 is larger than the first predetermined value A and smaller than the second predetermined value B, the controller 26 determines that the current value can be learned, and the actual current in other ranges. It is determined that the current value is not a learnable current value. When it is determined that the actual current of the fuel cell stack 1 is not a learnable current value, the controller 26 determines that the fuel cell stack 1 is not in a state where the current voltage characteristic learning process can be performed, and a series of learning is performed. The determination process ends. On the other hand, when it is determined that the actual current of the fuel cell stack 1 is a learnable current value, the controller 26 advances the learning execution determination process to the process of step S35.

  In the process of step S35, the controller 26 calculates the absolute value of the difference value between the actual current of the fuel cell stack 1 and the actual current when the previous learning process was performed, and determines whether or not the absolute value is greater than or equal to a predetermined value. Determine. Specifically, as shown in FIG. 13, the fuel cell stack 1 in an output current region in which the relationship between the reference voltage and the actual voltage difference ΔPMIV with respect to the output current of the fuel cell stack 1 cannot be expressed by the learning formula shown in the above formula 1. When driving, there is a possibility that learning is continuously performed in a narrow driving region, and erroneous learning is performed.

  Therefore, the controller 26 calculates the absolute value of the difference value between the actual current of the fuel cell stack 1 and the actual current value when the previous learning process is performed. If the absolute value is not equal to or greater than the predetermined value, the fuel cell stack 1 It is determined that the voltage characteristic learning process cannot be performed, and the series of learning execution determination processes ends. On the other hand, if the absolute value is greater than or equal to the predetermined value, the controller 26 advances the learning execution determination process to the process of step S36.

  In the process of step S36, the controller 26 sets the actual current of the fuel cell stack 1 to the learned output current. Thereby, the process of step S36 is completed and the learning execution determination process proceeds to the process of step S37.

  In the process of step S37, the controller 26 sets the value of the learning execution permission flag to 1, and ends a series of learning execution determination processes.

[Target current calculation processing]
The flowchart shown in FIG. 14 starts when the process of step S2 is completed, and the target current calculation process proceeds to the process of step S41.

  In the process of step S41, the controller 26 calculates the reference value of the target current corresponding to the required generated power calculated by the process of step S1, using the basic current-voltage characteristics of the fuel cell stack 1. More specifically, in this embodiment, the controller 26 requests the required power generation obtained by the processing in steps S1 and S22 from the map showing the relationship between the required generated power for each temperature and the target current basic value as shown in FIG. By reading the target current basic value corresponding to the power and temperature, the target current basic value of the fuel cell stack 1 is calculated.

  Note that the map shown in FIG. 15 shows the operating temperature using the characteristic data corresponding to the median value of the characteristic variation due to individual differences in the current-voltage characteristic at the beginning of operation, for example, as the reference current-voltage characteristic of the fuel cell stack 1. It is obtained by setting the current value for the generated power every time. Thereby, the process of step S41 is completed, and the target current calculation process proceeds to the process of step S42.

  In the process of step S42, the controller 26 calculates a correction value of the target current basic value using the learning parameters A (k) and B (k) obtained by the process of step S26. The correction value of the target current basic value is used to correct the target current in order to obtain the required generated power when the current-voltage characteristics of the fuel cell stack 1 change due to individual differences or changes over time. The current / voltage characteristics of the fuel cell stack 1, that is, the learning parameters A (k) and B (k) and the required generated power can be calculated. In this embodiment, the controller 26 is shown in FIG. A plurality of maps showing the relationship between the required generated power for each learning parameter value and the target current correction value are prepared, and the target current correction value is calculated by referring to this map.

  The map shown in FIG. 16 is used to calculate a correction value for the target current based on the learning parameter A (k) and the required generated power when the value of the learning parameter B (k) is fixed to a predetermined value. In order to obtain the required generated power when the actual current-voltage characteristic changes with respect to the current-voltage characteristic serving as the reference of the fuel cell stack 1 used in the setting of the map shown in FIG. The correction value of the current with respect to the basic value of the target current calculated by the above is calculated in advance and data is set.

  Also, a plurality of maps shown in FIG. 16 are prepared according to the value of the learning parameter B (k), and a target current correction value calculated in advance is set for each value of the learning parameter B (k). Then, the current correction value is calculated with reference to the plurality of maps, and the target current correction value is calculated by performing linear interpolation of the plurality of current correction values based on the value of the learning parameter B (k). Thus, by referring to a plurality of maps and interpolating the results, the correction value of the target current can be calculated based on the learning parameters A (k) and B (k) and the required generated power. Thereby, the process of step S42 is completed and the target current calculation process proceeds to the process of step S43.

  In the process of step S43, the controller 26 calculates the target current by adding the target current basic value calculated by the process of step S41 and the correction value calculated by the process of step S42. Thereby, the process of step S43 is completed and a series of target current calculation processes are complete | finished.

[Gas supply control processing]
The flowchart shown in FIG. 17 starts when the process of step S3 is completed, and the gas supply control process proceeds to the process of step S51.

  In the process of step S51, the controller 26 is calculated by the process of step S3 from the map showing the relationship between the target generated current and the target gas pressure as shown in FIG. 18 created in consideration of the power generation efficiency of the fuel cell stack 1. The target gas pressure is calculated by reading the target gas pressure corresponding to the target generated current. Thereby, the process of step S51 is completed and the gas supply control process proceeds to the process of step S52.

  In the process of step S52, the controller 26 controls the hydrogen pressure control valve 3 based on the target gas pressure calculated by the process of step S51. The hydrogen pressure control valve 3 is controlled by feedback controlling the command opening based on the deviation between the hydrogen pressure detected by the pressure sensor 19 and the target gas pressure.

  This feedback control can be performed by a generally well-known method such as PI control or model reference control. The command opening of the hydrogen pressure control valve 3 is input from the controller 26 to the drive circuit of the hydrogen pressure control valve 3, and the hydrogen pressure control valve 3 is driven according to the command opening. Thereby, the process of step S52 is completed, and the gas supply control process proceeds to the process of step S53.

  In the process of step S53, the controller 26 controls the flow rate of air supplied to the fuel cell stack 1 by controlling the compressor 10. Specifically, the controller 26 sets the target power generation current and the target air flow rate as shown in FIG. 19 which are set so as to achieve an air utilization rate that does not cause local shortage of air supply inside the fuel cell stack 1. The target air flow rate is calculated by reading the target air flow rate corresponding to the target generated current calculated by the process of step S3 from the map showing the relationship.

  Next, the controller 26 is based on a map showing the relationship between the target air flow rate for each target gas pressure and the compressor command rotational speed as shown in FIG. 20 set based on the air flow rate characteristics with respect to the rotational speed and pressure ratio of the compressor 10. The compressor command rotational speed is calculated by reading the compressor command rotational speed corresponding to the calculated target air flow rate. Then, the controller 26 drives the compressor 10 at the command rotational speed by outputting the calculated compressor command rotational speed to the drive circuit of the compressor 10. Thereby, the process of step S53 is completed, and the gas supply control process proceeds to the process of step S54.

  In the process of step S54, the controller 26 controls the air pressure control valve 12 based on the target air pressure calculated by the process of step S51. The air pressure control valve 12 is controlled by feedback controlling the command opening based on the deviation between the air pressure detected by the pressure sensor 21 and the target air pressure. The command opening of the air pressure control valve 12 is input from the controller 26 to the drive circuit of the air pressure control valve 12, and the air pressure control valve 12 is driven according to the command opening. Thereby, the process of step S54 is completed and a series of gas supply control processes are complete | finished.

  As is apparent from the above description, in the fuel cell system according to the first embodiment of the present invention, it is necessary for the controller 26 to output the required generated power based on the output characteristics serving as the reference of the fuel cell stack 1. The amount of current taken out from the correct fuel cell is calculated, and the relationship between the output characteristics serving as the reference of the fuel cell stack 1 and the actual output characteristics of the fuel cell stack 1 is learned. Then, the controller 26 calculates a correction amount of the extraction current amount based on the learning result, corrects the extraction current amount based on the calculated correction amount, and sets the corrected extraction current amount as the target current value as the fuel cell stack. The flow rate and pressure of hydrogen and air supplied to 1 are controlled.

  That is, in the fuel cell system according to the first embodiment of the present invention, the controller 26 learns the change in the output characteristics of the fuel cell stack 1 and corrects the extraction current amount of the fuel cell stack 1 based on the learning result. . Therefore, according to the fuel cell system according to the first embodiment of the present invention, the required power is accurately converted into the amount of current drawn from the fuel cell stack 1 with a simple configuration, and the power generation amount of the fuel cell stack 1 is accurately calculated. It can be controlled well.

  In addition, according to the fuel cell system according to the first embodiment of the present invention, the controller 26 determines the fuel cell stack 1 from the current-voltage characteristics that serve as a reference for the fuel cell stack 1 based on the operating state of the fuel cell stack 1. Since the reference voltage is calculated and the output current of the fuel cell stack 1 is input and the difference ΔPMIV between the reference voltage and the output voltage of the fuel cell stack 1 is output, the input / output relationship is learned. Thus, the change in the output of the fuel cell stack 1 can be easily detected.

  Further, according to the fuel cell system according to the first embodiment of the present invention, the reference current-voltage characteristic of the fuel cell stack 1 is the reference voltage of the fuel cell stack 1 with the output current and temperature of the fuel cell stack 1 as inputs. Since the controller 26 calculates the reference voltage of the fuel cell stack 1 based on the output current and temperature of the fuel cell stack 1, even if the output current or temperature changes. The relationship between the reference voltage and the output voltage can be learned.

  Further, according to the fuel cell system of the first embodiment of the present invention, the controller 26 receives the output current of the fuel cell stack 1 as an input and outputs the difference ΔPMIV between the reference voltage and the output voltage of the fuel cell stack 1 as an output. Since the input / output relationship is learned by using a learning expression composed of a linear function, the learning logic can be simplified, and a function with a small change in learning value with respect to a change in output current can be used. As a result, even if the operating conditions are limited, the learning accuracy can be improved using simple learning logic.

  Further, according to the fuel cell system according to the first embodiment of the present invention, the controller 26 stores the correction amount of the extraction current amount calculated in advance, and refers to the stored correction amount data. Since the correction amount of the current value is calculated, it is not necessary to calculate a solution of a higher order equation or perform a convergence calculation, and the calculation load can be reduced.

  Further, according to the fuel cell system of the first embodiment of the present invention, the controller 26 receives the output current of the fuel cell stack 1 as an input and outputs the difference ΔPMIV between the reference voltage and the output voltage of the fuel cell stack 1 as an output. Since the correction amount of the extraction current amount is calculated based on the coefficient of each term of the linear function, the correction amount of the extraction current amount can be easily calculated.

  Further, according to the fuel cell system of the first embodiment of the present invention, the controller 26 receives the output current of the fuel cell stack 1 as an input and outputs the difference ΔPMIV between the reference voltage and the output voltage of the fuel cell stack 1 as an output. The correction amount of the extraction current amount calculated in advance based on the coefficient of each term of the linear function is stored, and the correction amount of the extraction current value is calculated with reference to the stored correction amount data. It is not necessary to calculate the solution of the equation or perform the convergence calculation, and the calculation load can be reduced.

  The fuel cell system according to the second embodiment of the present invention has the same configuration as the fuel cell system according to the first embodiment, and the flow of target current calculation processing is the fuel cell according to the first embodiment. It is different from that of the system. Therefore, only the flow of the target current calculation process will be described below with reference to the flowchart shown in FIG. 21, and the description of the configuration and other processes will be omitted.

[Target current calculation processing]
The flowchart shown in FIG. 21 starts when the process of step S2 is completed, and the target current calculation process proceeds to the process of step S61. Since the processing in steps S61 and S62 has the same contents as the processing in steps S41 and S42 shown in FIG. 14, the description thereof is omitted below, and the description starts from the processing in step S63.

  In the process of step S63, the controller 26 detects the actual generated power of the fuel cell stack 1 so that the actual generated power of the fuel cell stack 1 becomes equal to the required generated power, and is calculated by the process of step S62. Change the correction value of the target current. Note that the actual generated power of the fuel cell stack 1 can be detected by multiplying the current value detected by the current sensor 24 by the voltage detected by the voltage sensor 23. The correction value of the target current can be changed based on the difference in generated power.

  Also, the amount of change is feedback control that converges a certain state to a target value, such as a method of calculating from the sum of the amount of change proportional to the power difference and the amount of change proportional to the integrated value of the power difference (so-called PI control logic). It can be calculated by logic. Thereby, the process of step 63 is completed, and the target current calculation process proceeds to the process of step S64.

  In the process of step S64, the controller 64 learns the change amount calculated by the process of step S63. The amount of change can be learned by using a general learning logic, such as a learning formula using the operating conditions of the fuel cell stack 1 such as the operating temperature and the generated current as input and the changed value as output. Can be used. Thereby, the process of step S64 is completed and the target current calculation process proceeds to the process of step S65. In addition, since the process of step S65 is the same as the process of step S43 shown in FIG. 14, the description is abbreviate | omitted below.

  As is apparent from the above description, according to the fuel cell system according to the second embodiment of the present invention, the controller 26 requires the required generated power so that the required generated power and the generated power of the fuel cell stack 1 become equal. Since the amount of correction of the extraction current amount is adjusted based on the difference between the generated power and the generated power, the extraction current amount can be accurately calculated in accordance with the actual output characteristics of the fuel cell stack 1.

  Further, according to the fuel cell system according to the second embodiment of the present invention, the controller 26 determines the difference between the required generated power and the generated power so that the required generated power is equal to the generated power of the fuel cell stack 1. Since the correction amount of the extraction current amount is learned based on this, the extraction current amount can be calculated accurately in a short time in accordance with the actual output characteristics of the fuel cell stack 1.

  As mentioned above, although embodiment which applied the invention made by the present inventors was described, this invention is not limited by the description and drawing which make a part of indication of this invention by this embodiment. That is, it should be added that other embodiments, examples, operation techniques, and the like made by those skilled in the art based on the above-described embodiments are all included in the scope of the present invention.

It is a schematic diagram which shows the structure of the fuel cell system used as the 1st and 2nd embodiment of this invention. It is a functional block diagram which shows the structure of the fuel cell system which concerns on this invention. It is a flowchart figure which shows the flow of the generated electric power control process used as the 1st and 2nd embodiment of this invention. It is a flowchart figure which shows the flow of the required electric power generation calculation process which becomes the 1st and 2nd embodiment of this invention. It is a map figure which shows the relationship between the amount of accelerator operation for every vehicle speed, and request | required generated electric power. It is a flowchart figure which shows the flow of the current voltage characteristic learning process used as the 1st and 2nd embodiment of this invention. It is a map figure which shows the relationship between the output current for every temperature of a fuel cell stack, and a reference voltage. It is a figure which shows the relationship between the output current of a fuel cell stack, and the difference value of a reference voltage and an actual voltage. It is a figure which shows the numerical formula used in the successive least squares method. It is a figure which shows the time change of the learning parameter accompanying a learning process. It is a flowchart figure which shows the flow of the learning implementation determination process used as the 1st and 2nd embodiment of this invention. It is a figure which shows a very low load area | region and a very high load area | region. It is a figure which shows the output current area | region which cannot express the relationship between the output current of a fuel cell stack, and the difference value of a reference voltage and an actual voltage with a learning formula. It is a flowchart figure which shows the flow of the target electric current calculation process used as the 1st Embodiment of this invention. It is a figure which shows the relationship between the request | required power generation for every temperature of a fuel cell stack, and a target current basic value. It is a figure which shows the relationship between the request | required power generation for every learning parameter, and a target electric current correction value. It is a flowchart figure which shows the flow of the gas supply control process used as the 1st and 2nd embodiment of this invention. It is a map figure which shows the relationship between target electric power generation current and target gas pressure. It is a map figure which shows the relationship between a target electric power generation current and a target air flow rate. It is a map figure which shows the relationship between the target air flow rate for every target gas pressure, and compressor command frequency | count. It is a flowchart figure which shows the flow of the target electric current calculation process used as the 2nd Embodiment of this invention.

Explanation of symbols

1: Fuel cell stack 2: High pressure hydrogen tank 3: Hydrogen pressure control valve 4: Hydrogen supply flow path 5: Hydrogen circulation flow path 6: Ejector 7: Hydrogen discharge flow path 8: Hydrogen purge valve 9: Air supply flow path 10: Compressor 11: Air discharge passage 12: Air pressure control valve 13: Coolant passage 14: Coolant circulation pump 15: Heat exchangers 16, 18, 22: Temperature sensors 17, 19, 21: Pressure sensor 20: Flow rate sensor 23: Voltage sensor 24: Current sensor 25: Power control device 26: Controller

Claims (9)

A fuel cell system having a fuel cell that generates power upon receiving a reaction gas,
A required generated power calculating means for calculating required generated power for the fuel cell;
Extraction current calculation means for calculating the amount of extraction current required from the fuel cell to output the required generated power based on the output characteristics serving as a reference of the fuel cell;
Output characteristic learning means for learning a relationship between an output characteristic serving as a reference of the fuel cell and an actual output characteristic of the fuel cell;
An extraction current correction amount calculating means for calculating a correction amount of the extraction current amount based on a learning result of the output characteristic learning means, and correcting the extraction current amount based on the calculated correction amount;
A fuel cell system comprising: control means for controlling the amount of reaction gas supplied to the fuel cell using the extraction current amount corrected by the extraction current correction amount calculation means as a target current value. The fuel cell system according to claim 1,
Current detecting means for detecting an output current of the fuel cell;
Voltage detecting means for detecting the output voltage of the fuel cell,
The output characteristic learning means includes
A reference voltage calculation means for calculating a reference voltage of the fuel cell from a current-voltage characteristic which is a reference of the fuel cell based on the operating state of the fuel cell;
Voltage difference learning means for learning an input / output relationship when an output current detected by the current detection means is used as an input and a difference between the reference voltage and the output voltage detected by the voltage detection means is output. A fuel cell system. The fuel cell system according to claim 2, wherein
The reference current-voltage characteristic of the fuel cell is a current-voltage characteristic in which an output current and temperature of the fuel cell are input and a reference voltage of the fuel cell is output, and the reference voltage calculation means is configured to output the output current of the fuel cell. A fuel cell system that calculates a reference voltage of the fuel cell based on temperature and temperature. The fuel cell system according to claim 2 or 3, wherein
The voltage difference learning means is input by using a learning equation consisting of a linear function having an output current detected by the current detection means as an input and a difference between the reference voltage and the output voltage detected by the voltage detection means as an output. A fuel cell system characterized by learning an output relationship. The fuel cell system according to any one of claims 1 to 4, wherein
The extraction current correction amount calculation means stores a correction amount of an extraction current amount calculated in advance, and calculates a correction amount of an extraction current value by referring to the stored correction amount data. Battery system. The fuel cell system according to any one of claims 2 to 4, wherein
The extraction current correction amount calculation means uses the output current detected by the current detection means as an input and outputs a difference between the reference voltage and the output voltage detected by the voltage detection means as a coefficient of each term of a linear function. A fuel cell system characterized in that a correction amount of an extraction current amount is calculated on the basis thereof. The fuel cell system according to any one of claims 2 to 4, wherein
The extraction current correction amount calculation means uses the output current detected by the current detection means as an input and outputs a difference between the reference voltage and the output voltage detected by the voltage detection means as a coefficient of each term of a linear function. A fuel cell system characterized in that a correction amount of an extraction current amount calculated in advance is stored, and a correction amount of the extraction current amount is calculated with reference to the stored correction amount data. The fuel cell system according to any one of claims 1 to 7,
Generated power detection means for detecting generated power of the fuel cell;
Means for adjusting a correction amount of the extraction current amount based on a difference between the required generated power and the generated power so that the required generated power and the generated power detected by the generated power detecting means are equal to each other. Fuel cell system. The fuel cell system according to claim 8, wherein
The extraction current correction amount calculating means calculates a correction amount of the extraction current amount based on a difference between the required generated power and the generated power so that the required generated power is equal to the generated power detected by the generated power detection means. A fuel cell system comprising means for learning.

Images