JP2008010176A - Abnormality diagnostic system and abnormality diagnostic method of fuel cell power generation system, fuel cell power generation system, and its operation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain necessary detection information even if monitor portions are reduced in the abnormality diagnosis of a fuel cell power generation system. <P>SOLUTION: Existence or nonexistence of the abnormality in a fuel cell stack or a power generation system containing the fuel cell stack is diagnosed from the magnitude of frequency component in a prescribed frequency range abstracted from voltage detected from a detection part using at least one cell constituting parts of the fuel cell stack. The cell forming the detection part is selected based on the intensity before shipment in the prescribed frequency range. Abnormality is accurately detected before it is grown. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an abnormality diagnosis system and an abnormality diagnosis method in a power generation system having a fuel cell stack. The present invention also relates to a fuel cell power generation system including an abnormality diagnosis unit and an operation method thereof. The present invention is suitable for application to PEFC (Polymer Electrolyte Fuel Cell).

  The conventional technology for abnormality diagnosis of fuel cell stacks is to divide a large number of cells that make up the fuel cell stack into multiple cell groups, measure the output voltage for each cell group, and calculate the average output voltage and voltage variation And the method of diagnosing an operation state based on the value of the output voltage for every cell group is known (for example, refer to patent documents 1).

JP 2004-127915 A (summary)

  According to Patent Document 1, voltage monitoring over the entire fuel cell stack is possible. However, in a practical fuel cell power generation system, there is a demand for simplification of the system, and it is desired to reduce the number of monitor points. If the number of monitor locations is reduced, there is a problem that information to be detected is lost. Therefore, in the abnormality diagnosis of the fuel cell power generation system, the problem is how to achieve both reduction in the number of monitor points and appropriate securing of information to be detected.

  An object of the present invention is to provide an abnormality diagnosis system for a fuel cell stack, an abnormality diagnosis method therefor, and a fuel cell power generation system provided with an abnormality diagnosis means, so that necessary detection information can be obtained even if the number of monitors is reduced. It is to provide a driving method.

  The present invention uses at least one cell that forms part of the fuel cell stack as a detection unit, extracts a frequency component in a predetermined frequency range determined in advance from the voltage detected by the detection unit, and from the magnitude of the frequency component, An abnormality diagnosis system for diagnosing the presence or absence of abnormality of the fuel cell stack or a power generation system including the fuel cell stack, wherein the cell serving as the detection unit is based on the strength before shipment in the predetermined frequency range. The abnormality diagnosis system in the fuel cell power generation system is characterized by being selected.

  The present invention also relates to a method for diagnosing an abnormality in a power generation system including a fuel cell stack, and detects a voltage using at least one cell selected based on a pre-shipment strength in a predetermined frequency range as a detection unit. The frequency component in the predetermined frequency range is extracted from the voltage, and the presence or absence of abnormality of the fuel cell stack or the power generation system including the fuel cell stack is diagnosed from the magnitude of the extracted frequency component. An abnormality diagnosis method characterized by the above.

Further, the present invention provides a fuel cell power generation system that includes a fuel cell stack and generates electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell.
Voltage detection means for detecting a voltage using at least one cell selected on the basis of a pre-shipment intensity in a predetermined frequency range as a detection unit;
Frequency component extraction means for extracting a frequency component in the predetermined frequency range from the voltage detected by the voltage detection means;
Amplifying means for amplifying the frequency component extracted by the frequency component extracting means;
A fuel cell power generation system comprising signal processing means for diagnosing the presence or absence of abnormality by comparing the signal intensity amplified by the amplification means with a threshold value obtained by a function correlating normal time and abnormal time is there.

  Further, the present invention is a method for operating a fuel cell power generation system having a fuel cell stack and generating an electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell. The voltage is detected using at least one cell selected on the basis of the strength before shipment in the frequency range as a detection unit, and a frequency component in a predetermined frequency range is extracted from the detected voltage, and the size of the extracted frequency component The fuel cell power generation system is characterized in that the presence or absence of abnormality of the fuel cell stack or the power generation system including the fuel cell stack is diagnosed.

  According to the present invention, the fuel cell stack is divided into a plurality of cell groups, the output voltage for each cell group is measured, and the abnormality of the cell stack and the fuel cell system can be determined without examining the average voltage and voltage variation. It is possible to detect with high accuracy even when the occurrence of abnormality is small.

  The present invention can further include the following embodiments.

  (A) At least two frequency ranges included in a predetermined frequency range determined in advance from the voltage detected by the detection unit, with at least one cell selected on the basis of the strength before shipment in a predetermined frequency range as a reference. An abnormality diagnosis system that extracts the frequency component of the fuel cell and diagnoses the presence or absence of abnormality of the fuel cell stack or the power generation system including the fuel cell stack from the magnitudes of at least two frequency components.

  (B) At least two frequency ranges included in a predetermined frequency range predetermined from the voltage detected by the detection unit, with at least one cell selected on the basis of the strength before shipment in a predetermined frequency range as a reference. An abnormality diagnosis that extracts a frequency component of the fuel cell, diagnoses an abnormality of the fuel cell stack from at least one of the extracted frequency components, and diagnoses an abnormality of the hydrogen production apparatus in the fuel cell power generation system from at least one of the other frequency components system.

  (C) A voltage is detected by using at least one cell forming a part of the fuel cell stack as a detection unit, a frequency component in a predetermined frequency range is extracted from the voltage, and the fuel cell is calculated from the size of the extracted frequency component. An abnormality diagnosis method for diagnosing the presence or absence of an abnormality in a power generation system including a cell stack or a fuel cell stack, wherein a cell serving as a voltage detection unit is selected and selected based on a strength before shipment in the predetermined frequency range. An abnormality diagnosis method for a fuel cell power generation system that performs abnormality diagnosis by a cell, determines whether or not to update a threshold value of the intensity of the predetermined frequency range during operation of the fuel cell, and maintains or updates the threshold value.

  (D) Extracting a frequency component in a predetermined frequency range from a voltage detected by using at least one cell forming a part of the fuel cell stack as a detection unit, and then calculating the fuel cell stack or fuel from the size of the extracted frequency component An abnormality diagnosis method for diagnosing presence / absence of an abnormality in a power generation system including a battery cell stack, wherein a cell serving as a voltage detection unit is selected on the basis of intensity before shipment in the predetermined frequency range, and abnormality diagnosis is performed using this cell. An abnormality diagnosis method for a fuel cell power generation system that starts and determines whether or not a detection unit should be updated during operation of the fuel cell, and continues or updates a cell serving as the detection unit.

  (E) A method of operating a fuel cell power generation system having a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell, and having a predetermined frequency range A voltage is detected using at least one cell selected on the basis of the strength before shipment as a detection unit, a frequency component in a predetermined frequency range is extracted from the voltage, and the fuel cell is calculated from the size of the extracted frequency component. An operation method of a fuel cell power generation system that diagnoses whether or not a cell stack or a power generation system including the fuel cell stack includes an abnormality, and changes a current value taken out from the fuel cell stack when the abnormality is detected.

  (F) A method of operating a fuel cell power generation system having a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell, and having a predetermined frequency range A voltage is detected using at least one cell selected on the basis of the strength before shipment as a detection unit, a frequency component in a predetermined frequency range is extracted from the voltage, and the fuel cell is calculated from the size of the extracted frequency component. First, a diagnosis is made of whether there is an abnormality in a cell stack or a power generation system including the fuel cell stack, and if an abnormality is detected, a current value taken out from the fuel cell stack is decreased until no signal indicating abnormality is output. Operation of the fuel cell power generation system that performs a control step and then performs a second control step that avoids the abnormality based on the cause of the abnormality Law.

  In the present invention, the frequency component in the predetermined frequency range is a signal that appears characteristically or significantly appears in the voltage signal detected by the voltage detector when an abnormality to be detected occurs in the fuel cell power generation system or the cell stack. Means a voltage signal component in a frequency range in which becomes larger.

  In the present invention, instead of dividing the entire fuel cell stack into monitor cell groups, at least a part of the cells constituting the fuel cell stack is used as a voltage detector (voltage monitor) to simplify the structure required for diagnosis. Yes. The cell serving as the voltage detection unit is selected based on the strength before shipment in a predetermined frequency range used for diagnosing the presence or absence of abnormality.

  The strength before shipment refers to the operation test performed before entering the actual operation, such as the strength obtained in the test operation performed before delivery of the fuel cell power generation system to the customer or the test operation performed after delivery to the customer. Means the strength obtained in

  Not only the average voltage of the voltage detector and the variance of the entire vibration component, but also the magnitude of the frequency component belonging to a specific frequency range set in advance is extracted from the detected voltage signal and diagnosed, so that the occurrence of abnormality is small It will be possible to detect from home accurately. Thereby, information loss to be detected can be prevented.

  An apparatus for performing abnormality diagnosis includes a voltage detection unit that detects a voltage from a cell serving as a voltage detection unit, and a frequency component extraction unit that extracts a frequency component in a predetermined frequency range determined in advance from the voltage detected by the voltage detection unit. Amplifying means for amplifying the frequency component extracted by the frequency component extracting means, and signal processing for diagnosing the presence or absence of abnormality by comparing the amplified signal intensity with a threshold value obtained by a function correlating normal time and abnormal time It is desirable to include means.

  It is desirable to provide the voltage detector in the separator of the fuel cell stack. By providing a detection terminal necessary for voltage detection in the separator of the fuel cell stack in advance, voltage detection for diagnosis can be performed without changing the size of the cell stack in the stacking direction.

  The number of cells included in the detection unit may be one, but a plurality of cells or all cells may be used as the detection unit.

  When extracting frequency components in at least two frequency ranges from the voltage detection unit and diagnosing the presence or absence of abnormality from the magnitudes of these frequency components, the following effects can be obtained. For example, by extracting the fluctuation (vibration component) of the entire stack voltage from the detection signal, and extracting and capturing the size of a plurality of frequency components specific to the fuel cell system or cell stack, for example, Comparison and confirmation of correlation are possible. This facilitates the separation of the background noise component and the signal belonging to the frequency region to be detected, and improves the accuracy of abnormality diagnosis. In addition, it is possible to diagnose an abnormality of the fuel cell stack from at least one of the extracted frequency components in at least two frequency ranges and to diagnose an abnormality of the hydrogen production apparatus from at least one of the other frequency components. The hydrogen production apparatus is provided in the fuel cell power generation system in order to produce hydrogen supplied as fuel gas to the fuel cell stack.

  In the present invention, the intensity of the predetermined frequency range used for diagnosing the presence or absence of abnormality can be updated as appropriate during operation. As a result, the influence of deterioration with time of the fuel cell stack or the like can be taken into the diagnosis.

  You may make it change a detection part on the basis of the intensity | strength updated at any time at the time of the driving | operation of the predetermined frequency range used in order to diagnose the presence or absence of abnormality. As a result, it is possible to detect an abnormality with higher accuracy.

  In the present invention, when an abnormality of the fuel cell stack or the power generation system is detected by diagnosis, it is desirable to change the current value extracted from the fuel cell by a predetermined amount by control. Along with the occurrence of abnormality, the amount of current taken out from the fuel cell stack is changed by a predetermined amount, so that the burden that leads to the deterioration of the cell stack can be quickly reduced and the operation of the power generation system can be continued stably.

  FIG. 1 is a schematic view showing an example of a fuel cell stack provided with abnormality diagnosis means. In the figure, reference numeral 100 is a fuel cell stack, 1 is a voltage detection unit for detecting a voltage Vd for diagnosis, 2a and 2b are voltage detection means (here, voltage detection terminals), and 3a and 3b are stacks. This is an output terminal for outputting the output voltage Vout. 4 is a DC component detection line for detecting a DC component of the diagnostic voltage Vd, and 5a and 5b are AC component detection lines for detecting a predetermined specific frequency component from the diagnostic voltage Vd. Reference numerals 6a and 6b denote specific frequency component extracting means for extracting specific frequency components, and here, bandpass filters are used. Reference numerals 7a and 7b denote amplification means (amplification circuit as an example) for amplifying the extracted minute frequency components, and reference numeral 8 denotes signal processing means (a processing unit using a microcomputer as an example) for abnormality diagnosis.

  As the specific frequency component extraction means 6a and 6b, a high-pass filter, a low-pass filter, and other signal filtering means can be used in addition to the band-pass filter. Specifically, it is only necessary to configure an electric circuit having an electrical characteristic that allows only a signal in the frequency domain to be extracted to pass significantly. Further, each frequency component may be extracted and amplified by performing a fast Fourier transform or other frequency analysis with a microcomputer or the like. Alternatively, the diagnostic voltage Vd may be differentiated so that frequency components other than direct current can be detected while the frequency components are very small. After the differentiation process, the corresponding frequency component may be extracted again.

  The specific frequency component extraction means 6a, 6b can be miniaturized as a dedicated integrated circuit together with the amplification means 7a, 7b and the signal processing means 8.

  The function of each part will be described. In the embodiment of FIG. 1, the voltage detection unit 1 is a set of cell stacks that form part of the fuel cell stack 100. In the case of the present embodiment, a set of cell stack parts is composed of about 10 cells. The voltage detector 1 is selected based on test data before shipment, but the voltage detector can be changed based on data during operation. The output voltage of the fuel cell stack 100 is taken out from the output terminals 3a and 3b, while the diagnostic voltage Vd is taken out from the voltage detection means 2a and 2b. The output terminal 3a and the voltage detection means 2a may be shared by the same terminal.

  By limiting the voltage detection unit 1 for diagnosis to a part of the fuel cell stack 100, the number of data to be processed during operation can be reduced. On the other hand, even if an abnormality occurs in cells other than the voltage detection unit 1, it can be detected by reviewing selected cells at arbitrary time intervals.

  FIG. 2 shows a case where the voltage detectors are set at two locations in the fuel cell stack. A voltage detector A indicated by reference numeral 1a detects a voltage Vd1 for diagnosis, and a voltage detector B indicated by 1b detects a voltage Vd2 for diagnosis. The voltage of the voltage detector A is detected by the voltage detectors 2a and 2b, and the voltage of the voltage detector B is detected by the voltage detectors 2c and 2d.

  In this example, it was confirmed from the measurement data before shipment that an abnormality is likely to occur at the positions of the voltage detection units A and B. When one of the voltage detection units A and B causes a voltage fluctuation exclusively depending on the operating state, that is, when one of the voltage detection units A and B does not occur in the other, the voltage detection units A and B What is necessary is just two places. When voltage fluctuations occur at the same time, either voltage detection unit A or B can be used as the voltage detection unit. FIG. 2 shows a case where there are two sets of voltage detection units, but the voltage detection position for detecting the frequency and the number of sets can be arbitrarily set.

  In the present invention, signal components in a predetermined specific frequency region are extracted, and abnormality diagnosis is performed based on the extracted signal components. If the fuel cell stack is normal, it outputs a substantially constant voltage depending on the operating conditions, but under unfavorable operating conditions, the voltage varies with time. At this time, when the voltage is developed in the frequency domain, there are many cases where components in a specific frequency domain are increased. Although there may be large variations in measurement, a significant correlation can be obtained with the occurrence of abnormality when looking at a certain frequency region. In the embodiment of FIG. 1, such a frequency region is extracted by means such as a band pass filter, and this is amplified and used for abnormality diagnosis processing. Of course, the signal component in the frequency domain may be extracted and processed by other means.

  When a specific frequency component having a significant correlation with the abnormality is extracted, other frequency components having no correlation are removed, so that the S / N ratio of the signal is improved. In addition, since the generation of minute frequency components is captured by the amplification means 7, the detection accuracy is improved. When looking at the entire fuel cell stack, it is rare that an abnormality occurs at a specific cell stack, but an abnormality does not occur at all in a specific cell. This is because the gas supply to the fuel cell stack flows to each cell through the manifold and the current is also taken out in series, so that the occurrence of abnormal cells extends to other cells regardless of the magnitude of the influence. Therefore, if there is a detection accuracy capable of accurately capturing this at the very initial stage where an abnormality has started to occur, a correct abnormality diagnosis can be performed even if the voltage detection unit is limited to the fuel cell stack portion.

  In the present invention, frequency components having a significant correlation with the occurrence of an abnormality are extracted and can be detected from a small component, so that the detection accuracy can be improved, and an avoidance process for the abnormality can be performed quickly.

  In detecting the diagnostic voltage Vd, it is desirable to keep the current value controlled and taken out from the fuel cell stack constant. This is because a change in current changes the diagnostic voltage Vd depending on the original characteristics of the fuel cell stack regardless of whether there is an abnormality. In system operation, if a predetermined current value is set for each predetermined operation state (operation mode) such as rated operation or partial load operation and controlled by the power controller, a substantially constant current control state can be obtained during system operation. Obtainable. Diagnosis may be performed in this state.

  To avoid diagnostic errors, the diagnostic processing by the signal processing means 8 is temporarily interrupted when starting or stopping, or when changing each operation mode (such as rated operation or partial load operation). You may make it do. Since the control means that controls the entire system can determine which operating state the system is in, the transient response period is determined from the sampled data by referring to the information related to the setting and switching of the operating state. It is possible to prevent the data corresponding to 1 from being diagnosed.

  The signal processing means 8 for diagnosis may be configured as a unit integrated with the control means for controlling the entire system, or may be configured as a separate unit to exchange information by mutual communication. .

  In the embodiment of FIG. 1, signal components in at least two frequency regions are extracted by two band pass filters (BPF1 and BPF2). When signal processing is performed using signals in two frequency domains as in this embodiment, at least the following two effects can be obtained.

  One is an effect when changes in frequency components caused by different causes of abnormalities appear as signal changes in different frequency regions. If the signals in each frequency domain are separated and extracted by BPF1 and BPF2 in FIG. 1, the cause of the abnormality can be specified by the difference in the frequency domain as well as detecting the abnormality with high accuracy.

  The other is an effect in the case where changes in frequency components caused by the occurrence of one abnormality appear simultaneously as signal changes in different frequency regions. By looking at the correlation between the two, it can be determined that an abnormality has occurred only when there is a signal change in any frequency region, so even if a noise component is superimposed on the signal component in that frequency region, it is accurate. Abnormal diagnosis is possible. As an example, when noise is superimposed on one frequency component among the signal components in the frequency domain, it is difficult to distinguish whether this is due to the occurrence of abnormality or noise due to the method of detecting only the frequency component. . However, if the signal in the frequency domain is separated and extracted, the signal component in the other frequency domain is confirmed at the same time, and if the signal component increases with correlation, it is determined that an abnormality has occurred. Otherwise, it can be determined that it is due to noise.

  Of course, in practice, the frequency to be detected may be only a signal component in one frequency region, and in this case, the configuration of the signal processing system for diagnosis can be simplified.

  Which frequency component is likely to cause an abnormality varies depending on the fuel cell stack, but if a frequency region where abnormality is likely to occur under certain operating conditions is specified in advance, abnormality diagnosis in the frequency region is possible. It is possible to separate the cause of the abnormality in the frequency range.

  FIG. 3 illustrates a method for setting a representative frequency to be detected. 10 to 17, the characteristic frequency band is blocked by a band pass filter like f1, f2, and f3 shown in the figure, and each signal intensity is compared with a normal value (actually, The difference is used), and the presence or absence of an abnormality is diagnosed in advance or based on a threshold set during operation. For example, the f1 block corresponds to a vibration component due to an abnormality in the fuel utilization rate, the f2 block corresponds to a vibration component due to an abnormality in the oxygen utilization rate, and the f3 block corresponds to a vibration component due to carbon monoxide (CO) mixing abnormality from a reformer or the like. Associate.

  In the above, the abnormality diagnosis mainly caused by the fuel cell stack has been described. However, the abnormality related to the reformer (hydrogen production device), the cooling water system, or the power conditioner should also be paid attention to the periodicity peculiar to the abnormality. And the same diagnosis can be made. Since the periodicity related to the cooling water system is often determined by the flow and thermal characteristics, it appears as a relatively long period. Also, abnormalities related to the inverter are often determined by electrical characteristics, and thus appear as a relatively short cycle.

  In addition to the above diagnosis, a DC component of the stack output voltage Vout and the diagnostic voltage Vd, that is, a component other than the component related to vibration may be used in the diagnosis.

  The magnitude of the voltage value is set as one of the abnormality determination conditions, and the presence / absence of abnormality is determined based on an AND condition that a specific frequency component appears in the voltage and a gradual change (drift) in Vout or Vd is indicated. You may do it.

  By making an abnormality diagnosis based on both a characteristic change appearing in a specific frequency component and a characteristic change appearing in a direct current component, the accuracy of diagnosis can be improved, and a more detailed cause can be easily identified.

  In addition, since the values of Vout and Vd show a gradual decrease over time even in a normal fuel cell stack, for example, based on the voltage value, a threshold value for abnormality diagnosis is called on a map. For example, a change in threshold value due to a change with time can be appropriately corrected, and an abnormality diagnosis including a change with time can be performed.

  FIG. 4 is a frame diagram for explaining a specific configuration example of the voltage detection unit 1. In the figure, reference numerals 101a and 101b denote end plates for maintaining the fuel cell stack under predetermined tightening conditions, reference numeral 102 denotes a separator, and reference numerals 103a, 103b, and 103c denote current collecting plates. Here, the output terminal 3a of the fuel cell stack and the voltage detection means 2a are the same as each other. In addition, electrolyte membranes, diffusion layers, sealing materials, cooling means, and the like that are not directly related to the description are omitted.

  The voltage detector 1 was configured as a cell stack between the current collector plate 103a and the current collector plate (voltage detection intermediate plate) 103b. Looking at the difference from the structure of the original fuel cell stack, it can be seen that the voltage detector according to the present invention can be configured with a small change that mainly inserts the current collector plate 103b, which is a voltage detecting intermediate plate. In addition, by using the current collector plate 103b, which is a voltage detection intermediate plate, it is possible to reduce the occurrence of wiring connection failure over time and to ensure voltage detection.

  FIG. 5 is another frame diagram for explaining a specific configuration example of the voltage detection unit 1. Here, the current collector plate 103b is not used, and the voltage detector 1 is configured as a cell stack between the voltage detectors 2a and 2b. Instead of using a special current collector plate, the voltage detection means 2a and 2b are provided integrally with the separator at the end of a normal separator.

  As a method of integration, there are the following methods. In the case of a molded separator or a metallic separator, a terminal for voltage detection can be integrally formed in advance at a position as shown in the figure of the separator. A terminal which is a voltage detection means may be provided with a structure suitable for a hole for electrical wiring, electrical wiring, or the like. Further, in the case of a dense graphite separator whose shape is relatively difficult to change, a hole capable of embedding a voltage detection terminal may be provided in the separator as a voltage detection means. Looking at the difference from the structure of the original fuel cell stack, the voltage detection means 2a and 2b are mainly integrated with a predetermined separator, so that the cell stack is compared with the case where a current collector plate is added. The dimension in the stacking direction of itself is not increased. Moreover, since the voltage detection means 2a, 2b can be integrated in advance with a predetermined separator, there is no need to change the stacking procedure of the fuel cell stack.

  FIG. 6 is a schematic diagram for explaining a diagnostic voltage detection unit of the fuel cell stack. The configuration of each part will be described. In the figure, 105a and 105b are integrated MEAs, and 106a and 106b are electrode portions. MEA is a membrane / electrode assembly, which is an integrated electrolyte membrane and electrode necessary for power generation of a fuel cell. The integrated MEA is not only an electrode but also a member located around the MEA, such as a gas diffusion layer and a sealing material (gasket). Here, an integrated MEA or the like is schematically shown. The number of separators and integral MEAs is limited to those necessary for the explanation, and only the other components necessary for the explanation are shown.

  Here, for the sake of simplicity, the case of using an integrated MEA is illustrated, but of course, a non-integrated MEA and other members may be used in combination.

  FIG. 7 shows an example of a cross-sectional structure of the integrated MEA. Reference numeral 502 in the figure denotes an MEA, which is a film. Reference numeral 503 denotes a gasket, which is disposed above and below the MEA and can be sealed so as to surround the periphery. Reference numeral 504 denotes a gas diffusion layer, which is arranged above and below the MEA so as to fit in the central portion other than the gasket 503. Reference numeral 501 denotes a manifold, which is a hole provided for passing gas or water. Since the hole is provided in the MEA 502 at the gasket 503 portion, the hole is sealed by the gasket 503 to prevent leakage of gas and water.

  Subsequently, the function of each unit will be described. In FIG. 6, the structure in which the cell stacking part is formed between the voltage detecting means 2a and 2b and this is used as the voltage detecting part 1 is the same as that of Example 1 according to the first embodiment. By devising the structure of at least one cell included in the detection unit 1, abnormality detection for diagnosis can be accurately performed.

  In FIG. 6, at least one cell of the voltage detection unit includes a membrane / electrode assembly (MEA) having a smaller electrode area than that of the power generation unit other than the detection unit. Specifically, the area of the electrode part 106a was made equal to that of the power generation part, whereas the area of the electrode part 106b was made smaller than that. Each flow path of the separator crosses all the electrode portions, but current concentrates in a limited area in the electrode portion 106b with respect to the electrode portion 106a. A cell having a small electrode area or electrode effective area is operated at a relatively high current density, so that the operating conditions are always stricter than those of other power generation units. As a result, an abnormality is likely to occur in the voltage detection unit prior to the power generation unit, so that the occurrence of the abnormality can be detected at an earlier stage than the cell of the power generation unit. In this way, the sensitivity of the voltage detector can be improved.

  In FIG. 6, the configuration example of the voltage detection unit 1 is shown. However, if the voltage detection means 2a and 2b are output terminals 3a and 3b, respectively, an electrode area and an electrode effective area are included in a normal fuel cell stack. It can also be regarded as a configuration in which small cells are provided. In this case, signal processing for abnormality diagnosis may be performed using the cell of the electrode unit 106b as a voltage detection unit.

  FIG. 8 shows an example in which an abnormality that occurs in a special cell provided in a part of the voltage detector or the fuel cell stack is detected by temperature. Reference numeral 108 denotes a temperature measuring means in which a thermocouple, a semiconductor type, and other temperature sensors are arranged in a thin tube shape. The tube is inserted into a hole provided in the separator 102. By adjusting the length of the tube, the position in the separator can be determined. If a temperature detector is provided at the tip of the tube, it is possible to adjust which part of the separator is to be measured. Abnormalities in the fuel cell stack appear not only in the voltage but also in the temperature change, so this is detected. In particular, an abnormal rise in temperature should be avoided for safety, and this is ensured by direct measurement of temperature. Abnormality detection based only on temperature is possible, but when combined with abnormality detection based on voltage, abnormality diagnosis is more reliable.

  As described above, in the voltage detection unit according to the second embodiment, the detection accuracy of the abnormal state can be increased by devising the internal structure of the voltage detection unit and the measurement method.

  As an example, the voltage detection unit according to the second embodiment further extracts a frequency component in a predetermined frequency range from the voltage detected by the voltage detection unit, and determines the fuel cell stack or the fuel cell from the magnitude of the frequency component. What is necessary is just to diagnose the presence or absence of an abnormality in the system including the cell stack. In this case, the voltage instability phenomenon of the voltage detection unit, which occurs at an earlier stage than the cell of the power generation unit, can be accurately detected by extracting characteristic frequency components, so that it is highly effective in early detection of abnormal conditions. In addition, the cause of the abnormality can be easily identified by separating the signal for each frequency region and performing the diagnostic processing.

  FIG. 9 shows an example in which the fuel cell power generation system according to the third embodiment is applied to a stationary distributed power source disposed in each home. Reference numeral 700 in the figure denotes a stationary distributed power source, which includes at least part of the fuel cell system according to the present invention. Operation as a cogeneration system was assumed.

  In the stationary distributed power source 700, in the reformer (hydrogen production device) of the fuel cell system, gas and air supplied from the outside, ion-exchanged water produced from pure water or tap water generated as a result of fuel cell power generation, etc. To produce hydrogen. As the raw material gas, natural gas mainly composed of methane or city gas can be used. Propane gas or other fuel may be supplied by a cylinder or the like. When using city gas, it is known that the sulfur component contained in the odorant poisons the catalyst, so it is supplied to the catalytic reaction section through a desulfurizer.

  A feature of using a fuel cell for a stationary distributed power supply is that it can provide not only power generation but also hot water obtained by exhaust heat from the fuel cell. In the case of a solid polymer combustion cell, the temperature during power generation is about 70 to 80 ° C., and the temperature inside the fuel cell is adjusted using cooling water or the like. Hot water is obtained by recovering the heat generated by the reaction of the fuel cell or internal resistance by cooling. However, if water supplied from the outside is used directly for cooling the fuel cell, it may adversely affect the fuel cell due to impurities contained in the water. In such a case, use means having a heat exchange function from the outside. What is necessary is just to heat up the water to supply indirectly. Since the heated hot water becomes, for example, about 50 to 60 ° C., if this hot water is stored and used in a hot water storage tank, the hot water used in the kitchen, bath or hand-washing can be provided instead of the water heater. In addition, in a fuel cell, the electric power obtained by electrochemically reacting an oxidizing gas such as air and a fuel gas such as hydrogen to generate electricity is combined with a variety of domestic electrifications together with externally supplied power. Since it can be used to drive products, the amount of power supplied from the outside can be reduced. Of course, if there is sufficient power generation capacity, power can be supplied without externally supplied power.

  When the temperature of the water supplied from the outside is low and the temperature rise is not sufficient, or when the water temperature in the hot water storage tank is lowered, a separate heating means may be provided. The heating means can raise the temperature of the water by, for example, burning part of the source gas supplied from the outside. The feed water temperature can be maintained at a predetermined temperature by feedback control for adjusting the heating amount and the flow rate of the hot water. A similar system may be configured in combination with a commercially available gas burner.

  When the fuel cell power generation system according to the present invention is applied to a stationary stationary power source for household use, in the abnormality diagnosis of the fuel cell and the fuel cell system, necessary detection information can be obtained even if the number of monitor points is reduced. With a compact configuration, the long-term life expected for a home system can be easily secured by early detection of abnormalities and appropriate avoidance control while keeping costs low.

  In the fuel cell system according to the present invention, the voltage change is accurately and promptly detected, and after abnormality diagnosis is performed, feedback control is performed to quickly avoid the abnormal state. The state can be restored to a normal state, and the characteristic deterioration of the entire fuel cell stack including the voltage detector can be suppressed. In particular, the voltage detection unit may employ an arrangement (an example of the first embodiment) or a configuration (second embodiment) that makes it easy to detect an abnormality. If feedback control that avoids the abnormal state is performed promptly, the abnormality that occurred in the voltage detection unit can be recovered to a normal state at a very small stage, so that significant deterioration in a short period can be eliminated. it can. That is, by combining appropriate feedback control, it is possible to suppress the characteristic deterioration of the detection unit while increasing the sensitivity of the abnormality detection unit. In this case, when the abnormality starts to occur in the voltage detection unit, almost no abnormality occurs in the other power generation units, so that there is no deterioration of the entire fuel cell stack.

  Here, specific control based on abnormality diagnosis includes the following methods. First, the diagnosis voltage Vd of the voltage detection unit is processed to diagnose whether or not an abnormality has occurred. When an abnormality is detected, the current control amount of the fuel cell stack 100 is quickly reduced. By lowering the current value, the fuel utilization rate of the fuel cell stack 100 is relatively lowered, so that the occurrence of an abnormality and the burden that leads to the deterioration of the cell stack are quickly reduced, and the system operation is continued stably. it can. As a result, a long-life system operation can be performed.

  The control related to the current amount may be performed for a predetermined time, and may be returned to the original current amount after the predetermined time. Or after the said control, you may make it diagnose that the abnormal state was avoided and return to the original electric current amount.

  In the above method, the reason for lowering the current control amount is as follows. In practical system operation, various abnormal states are assumed as shown in (1) to (4) below. (1) When the composition and flow rate of gas supplied from the reformer are not appropriate. (2) When water stays in the cell stack. (3) When the amount of gas leak (cross leak) through the electrolyte membrane increases. (4) The cell stack electrode catalyst is poisoned or deteriorated. On the other hand, in the case of (1), it is necessary to reduce the current because the net hydrogen supply amount is reduced. In the case of (2), the flow and diffusion of gas are hindered by water, and the cell It is desirable to reduce the current because a portion where hydrogen is not supplied locally is generated inside the stack. In the case of (3), it is necessary to reduce the current because the amount of hydrogen that effectively contributes to power generation is reduced. In the case of (4), the current is reduced because the catalyst that contributes to power generation is reduced. It is desirable. As described above, in many cases, by lowering the amount of current from the set value in the regulations, it is possible to quickly avoid an operation state that significantly deteriorates the cell stack.

  When an abnormality is detected, as the feedback control of the system, first, the current of the fuel cell stack is reduced by a predetermined amount in an emergency evacuation, and then individual feedback control for avoiding the abnormality is performed according to the diagnosed cause of the abnormality It is desirable to do.

  The feedback control for avoiding abnormality following the current control of the fuel cell stack specifically depends on the configuration of each system. For example, if the reformer temperature of the reformer is not appropriate, There is a method of changing the input raw fuel amount. If the heat recovery amount of the cooling water system is not appropriate, there is a method of changing the water circulation amount of the cooling water system.

  If the diagnosed abnormality is due to, for example, a reformer temperature abnormality, if the current of the fuel cell stack is lowered too much, the amount of hydrogen returned to the reformer will increase and the system operation will become unstable. May be. In such a case, the result of the abnormality diagnosis can be appropriately reflected in the current control so that the current decrease amount of the fuel cell stack is relatively reduced with the diagnosis result.

  On the other hand, when the combustor temperature of the reformer is too high, the amount of hydrogen returned to the combustor can be relatively decreased by increasing the amount of current, and the recovery of the heat balance can be promoted. Of course, if it is not desirable to change the amount of current, it may not be changed.

  In order to specify the cause of the abnormality in detail, information such as the temperature and pressure separately detected from the reformer, the auxiliary machine, and other parts may be referred to together with the abnormality diagnosis result. As an example, when a specific vibration frequency component accompanying a change in the reformed gas composition is generated in the voltage of the voltage detector, a CO concentration abnormality in the reformed gas is suspected. Therefore, first, the control for lowering the amount of current of the fuel cell stack is performed, and further referring to the temperature of the CO oxidation removal unit of the reformer, in the case where this is out of the proper temperature range, Judging that the reaction of the CO oxidation removal unit is no longer appropriate and increasing or decreasing the amount of air and water supplied to the CO oxidation removal unit in a direction to recover the abnormal state, stable operating state (normal state) Prompt to recover.

  In the above determination, information such as whether or not the DC component of the voltage of the voltage detector drifts may be further determined.

  As described above, in the stationary home-use distributed power source to which the fuel cell system according to the present invention is applied, when abnormality of the fuel cell stack or system is detected by diagnosis, the current value taken out from the fuel cell is controlled by the control. By changing the predetermined amount, it is possible to quickly reduce the load that leads to the deterioration of the cell stack, and to continue the operation of the system stably.

  In addition, the current amount can be set appropriately for each detected cause of abnormality. In the case where the current amount is controlled for a predetermined time, an appropriate value can be set for each detected abnormality cause for the length of the predetermined time.

  The fuel cell system and the control method according to the present invention may be applied to fuel cell systems other than those for home use. As in the case of home use, it is easy to continue stable operation by detecting abnormalities early and appropriate avoidance control while keeping the cost low with a compact structure. In this case, it is possible to quickly reduce the burden that leads to the deterioration of the fuel cell stack together with the detection of the abnormality, and the system operation can be continued stably.

  In this example, specific contents of the signal processing means 8 in the first embodiment will be described with reference to FIG. The fuel cell stack voltage acquired at a certain interval is read (801). Here, 2048 pieces are read. In order to extract a vibration component from the read voltage data, an average voltage of each cell is obtained (802). The average voltage is subtracted from each cell voltage (803). The voltage data is subjected to a fast Fourier transform (FFT) (804). The result is converted into a signal intensity spectrum (805). From this, trivial aliases and symmetric parts are removed (806). The block is divided into n (eight in this embodiment) blocks (frequency bands) (807). As a result, the frequency band can be extracted. FIG. 11 shows a schematic diagram of the intensity spectrum. When this processing is performed, a single scalar value can be assigned to each block. In addition to this processing, the same effect can be obtained by performing frequency extraction processing by hardware and by other methods. The processing from 801 to 807 is performed for all the measurement data numbers (808).

  Based on the data (intensity spectrum) subjected to this basic processing, the abnormality diagnosis processing shown in FIGS.

  The processing shown in 801 to 808 in FIG. 10 is performed on the data of each fuel cell or fuel cell block under normal operation and abnormal operation before shipment, and the signal strength under normal operation and abnormal operation is shown. A spectrum is obtained (809). An average signal strength of the power generation cells of the entire stack with respect to the normal operation is obtained, and a difference FFT spectrum with respect to the signal strength spectrum with respect to the abnormal operation is obtained (810). According to the diagnostic criteria (812), an abnormality diagnosis sensor cell is selected (811). In the diagnostic criteria (812), Uf represents an abnormal fuel utilization rate, Uo represents an abnormal oxygen utilization rate, and CO concentration represents an abnormal carbon monoxide. Carbon monoxide abnormality mainly occurs when there is an abnormality in the reformer. A characteristic coefficient matrix (or characteristic approximation coefficient matrix) that correlates the signal intensity FFT spectrum during normal operation and the signal intensity difference FFT spectrum during abnormal operation is obtained (813). Of the cells selected in 811, the abnormal operation signal strength difference FFT spectrum is estimated from the normal operation signal strength FFT spectrum by applying these matrices to the frequency blocks corresponding to various abnormalities shown in 815. A threshold value is determined from the value of the frequency block indicated at 815 (816). Based on this threshold, a map is created for all cells and all warning conditions (817). This map is first created by measuring the cell voltage in the fuel cell stack pre-shipment inspection. Further, by appropriately updating this abnormality diagnosis map during operation, it is possible to capture information including a history of changes in the fuel cell stack over time, etc. into the map, and a highly accurate fuel cell system Abnormal diagnosis is possible.

  13 and 16 show an abnormality diagnosis flow using the abnormality diagnosis map created in FIG. FIG. 13 is an abnormality diagnosis flow including determination of whether or not the threshold value needs to be updated, and FIG. 16 is an abnormality diagnosis flow including determination of whether or not the detection unit cell needs to be updated.

  An abnormality diagnosis flow in the case of including a determination whether or not to update the threshold value first will be described with reference to FIG. When the steps 801 to 807 in FIG. 12 are performed (818), it is confirmed by the counter check (819) whether or not the update interval has been reached. When the set number of times of voltage measurement is less than N, diagnosis (824) is performed using the threshold value in the abnormality diagnosis map. When the set number of times of voltage measurement is reached, it is determined whether the threshold needs to be updated (820).

  Whether or not to update the threshold value is determined by determining the size of the parameter shown in FIG. 15 (value obtained from the sensitivity and the determination range parameter) and determining whether there is an abnormality. When it is necessary to review the threshold value, the threshold value is adjusted (821). In the case of (M / L) <K−J, since it is insensitive to the set diagnostic sensitivity parameter, there may be many misdiagnosis, so the threshold value is adjusted (usually small) so that an abnormal diagnosis signal is output. In the case of K + J <(M / L), the threshold value is adjusted (usually increased) so as to reduce misdiagnosis because it is more sensitive than the set diagnostic sensitivity parameter. This process is performed according to the procedure shown in FIG.

  When an abnormal signal is generated by adjusting the threshold value, it is determined manually whether the abnormality diagnosis is correct or compared with data such as operating conditions (826). The number of occurrences of the abnormal signal is stored in the counter L (827). Further, the number of misdiagnosis is stored in the error counter M (828). Next, a diagnostic sensitivity parameter K is set (829). This parameter may be fixed from the beginning or may be reviewed during operation. Here, the sensitivity of the threshold update changes with 0.5 as the center, and the sensitivity becomes dull when it is close to 1, and the sensitivity becomes sensitive when it is close to zero. When the parameter is close to 1, the threshold is not reviewed even if there are many misdiagnosis, and when the parameter is close to zero, the threshold is frequently updated even if there are few misdiagnosis. Thereafter, an update determination range parameter J is set (829). Thereafter, the threshold value data for abnormality diagnosis is updated (822).

  If it is determined that there is no need to update the threshold data, the counter is reset as it is (823), and abnormality diagnosis (824) is performed. Finally, it is determined whether or not the operation has ended (825), and the continuation / end of the abnormality diagnosis routine is determined.

  Next, the abnormality diagnosis flow in the case where the determination as to whether or not to update the detection unit cell is performed will be described with reference to FIGS. 16 and 17.

  First, as in the case of FIG. 13, if the steps 801 to 808 are performed (918), a counter check (919) is performed, and if the set number of times of voltage measurement is less than N, the abnormality diagnosis map is displayed. Diagnosis (924) is performed using the inner threshold. When the set number of times of voltage measurement N is reached, it is determined whether or not the detection unit needs to be updated (920). This determination is performed according to the flow shown in FIG.

  First, the required time voltage is measured for all of the attached voltage detectors (926). For each voltage, FFT analysis or the like is performed in the procedure from 801 to 807 in FIG. 10 (927). Using the characteristic coefficient matrix (Uf, Uo, CO concentration), each predicted intensity spectrum is obtained (928). A detection unit that is considered to have the highest sensitivity, for example, a detection unit in which the intensity of the frequency range of interest is considered to easily exceed the threshold is selected (929). It is determined whether the selected detection unit is the same as the current detection unit (930).

  If the current detection unit is different from the current detection unit, the process returns to FIG. 16, and the threshold value of the new detection unit is cited from the threshold map (921). Next, the detection unit number and diagnostic threshold data are updated (922).

  When it is the same as the current detection unit and when the update of the detection unit is completed, the counter is reset (923), and abnormality diagnosis is performed (924). Finally, it is determined whether or not the operation has ended (925). If it is determined that the operation has ended, the diagnosis is ended.

  These diagnosis flows can be skipped when a separate abnormality diagnosis method is secured, such as when the fuel cell system is started / finished or when an operating parameter changes suddenly.

It is the schematic of the fuel cell stack provided with the abnormality diagnosis means. It is a perspective view of a fuel cell stack in which two voltage detection units are provided in the middle of the fuel cell stack. It is explanatory drawing which shows an example of the setting method of the typical frequency made into a detection target. It is a frame diagram explaining the specific structural example of a voltage detection part. It is another frame figure explaining the specific structural example of a voltage detection part. It is a schematic diagram which shows the voltage detection part for a diagnosis of a fuel cell stack. It is sectional drawing which showed an example of integrated MEA. It is explanatory drawing in the case of detecting the abnormality which arises in the special cell provided in the voltage detection part or a part of fuel cell stack by temperature. It is the schematic which shows the example at the time of applying a fuel cell electric power generation system to the stationary stationary power supply for household use. It is the flowchart which showed the detail of the frequency component in a voltage signal. It is a schematic diagram of a voltage vibration component intensity spectrum. It is the figure which showed the preparation flow of the diagnostic map in the measurement before driving | operation for diagnosing the abnormality resulting from a fuel usage rate, an oxygen usage rate, and carbon monoxide, and the update during driving | operation. It is the figure which showed an example of the abnormality diagnosis flow during a fuel cell system driving | operation. It is the figure which showed the example which judges whether it is necessary to implement the update of a threshold value. It is the figure which showed the flow which performs the review of a threshold value. It is the figure which showed the other example of the abnormality diagnosis flow during a fuel cell system driving | operation. It is the figure which showed the flow which judges whether the update of a detection part needs to be implemented.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Voltage detection part, 2a ... Voltage detection means, 2b ... Voltage detection means, 3a ... Output terminal, 3b ... Output terminal, 4 ... DC component detection line, 5a ... AC component detection line, 5b ... AC component detection line, 6a ... specific frequency component extracting means, 6b ... specific frequency component extracting means, 7a ... amplifying means, 7b ... amplifying means, 8 ... signal processing means, 100 ... fuel cell stack, 101a ... end plate, 101b ... end plate, 102 ... Separator, 103a ... current collector plate, 105a ... integrated MEA, 106a ... electrode section, 108 ... temperature measuring means, 501 ... manifold, 502 ... MEA, 503 ... gasket, 504 ... gas diffusion layer, 700 ... stationary distributed power source.

Claims (12)

  Using at least one cell forming a part of the fuel cell stack as a detection unit, a frequency component in a predetermined frequency range is extracted from a voltage detected by the detection unit, and the fuel cell from the magnitude of the frequency component An abnormality diagnosis system for diagnosing the presence or absence of an abnormality in a power generation system including a stack or the fuel cell stack, wherein the cell serving as the detection unit is selected based on the strength before shipment in the predetermined frequency range An abnormality diagnosis system in a fuel cell power generation system characterized by the above.   2. The fuel cell power generation according to claim 1, wherein the predetermined frequency range is a frequency range in which a frequency component characteristically appears in the voltage signal when an abnormality occurs or a frequency range in which the voltage signal specifically increases. Abnormality diagnosis system in the system.   2. The voltage detection unit according to claim 1, wherein the voltage detection unit detects a voltage from the detection unit, the frequency component extraction unit extracts a frequency component in the predetermined frequency range from the voltage detected by the voltage detection unit, and the frequency component extraction unit. Amplifying means for amplifying the extracted frequency component, and signal processing means for diagnosing the presence or absence of abnormality by comparing the amplified signal intensity with a threshold value obtained by a function correlating normal time and abnormal time, An abnormality diagnosis system for a fuel cell power generation system.   Of the cells forming the fuel cell stack, at least one cell selected on the basis of a pre-shipment strength in a predetermined predetermined frequency range is set as a detection unit, and the voltage detected by the detection unit is changed to the predetermined frequency range. The frequency components in at least two frequency ranges included are extracted, and the presence or absence of abnormality of the fuel cell stack or the power generation system including the fuel cell stack is diagnosed from the magnitude of the at least two frequency components. An abnormality diagnosis system in a fuel cell power generation system characterized by the above.   Of the cells forming the fuel cell stack, at least one cell selected on the basis of a pre-shipment strength in a predetermined predetermined frequency range is set as a detection unit, and the voltage detected by the detection unit is changed to the predetermined frequency range. Extract frequency components in at least two frequency ranges included, diagnose an abnormality in the fuel cell stack from at least one of the extracted frequency components, and detect an abnormality in the hydrogen production device in the fuel cell power generation system from at least one other An abnormality diagnosis system for a fuel cell power generation system, characterized in that a diagnosis is made.   In an abnormality diagnosis method for a power generation system including a fuel cell stack, at least one cell serving as a voltage detection unit is selected on the basis of a pre-shipment strength in a predetermined frequency range, and a voltage detected by the selected cell is used. A fuel characterized by extracting frequency components in the predetermined frequency range and diagnosing the presence or absence of abnormality of the fuel cell stack or the power generation system including the fuel cell stack from the magnitude of the extracted frequency components Abnormality diagnosis method for battery power generation system.   Using at least one cell forming a part of the fuel cell stack as a detection unit, a frequency component in a predetermined frequency range is extracted from a voltage detected by the detection unit, and the fuel cell from the magnitude of the frequency component An abnormality diagnosis method for diagnosing the presence or absence of an abnormality in a power generation system including a stack or the fuel cell stack, wherein a cell serving as a voltage detection unit is selected on the basis of strength before shipment in the predetermined frequency range, An abnormality diagnosis is performed with the selected cell, and whether or not the threshold value of the intensity in the predetermined frequency range should be updated during the operation of the fuel cell, and the threshold value is maintained or updated. Abnormal diagnosis method.   Using at least one cell forming a part of the fuel cell stack as a detection unit, a frequency component in a predetermined frequency range is extracted from a voltage detected by the detection unit, and the fuel cell from the magnitude of the frequency component An abnormality diagnosis method for diagnosing the presence or absence of an abnormality in a power generation system including a stack or the fuel cell stack, wherein a cell to be a detection unit is selected on the basis of strength before shipment in the predetermined frequency range, An abnormality diagnosis method for a fuel cell power generation system, wherein abnormality diagnosis is started by the cell, and whether or not the detection unit should be updated during operation of the fuel cell is determined and the cell serving as the detection unit is continued or updated . In a fuel cell power generation system comprising a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell,
Voltage detecting means for detecting a voltage using at least one cell selected on the basis of an intensity before shipment in a predetermined frequency range as a reference;
Frequency component extraction means for extracting a frequency component in the predetermined frequency range from the voltage detected by the voltage detection means;
Amplifying means for amplifying the frequency component extracted by the frequency component extracting means;
A fuel cell power generation system comprising signal processing means for diagnosing the presence or absence of an abnormality by comparing the signal intensity amplified by the amplifying means with a threshold value obtained by a function correlating normal time and abnormal time.   An operation method of a fuel cell power generation system having a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell, and is a pre-shipment in a predetermined frequency range A voltage is detected using at least one cell selected on the basis of intensity as a detection unit, a frequency component in a predetermined frequency range is extracted from the detected voltage, and the fuel cell stack is extracted from the size of the extracted frequency component Alternatively, a fuel cell power generation system operating method is characterized in that the presence or absence of abnormality in a power generation system including the fuel cell stack is diagnosed.   An operation method of a fuel cell power generation system having a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell, and is a pre-shipment in a predetermined frequency range A voltage is detected using at least one cell selected on the basis of intensity as a detection unit, a frequency component in the predetermined frequency range is extracted from the detected voltage, and the fuel cell stack or the A method for operating a fuel cell power generation system, comprising: diagnosing the presence or absence of an abnormality in a power generation system including a fuel cell stack and changing a current value taken out from the fuel cell stack when the abnormality is detected.   An operation method of a fuel cell power generation system having a fuel cell stack and generating electric energy by electrochemically reacting an oxidizing gas and a fuel gas in the cell, and is a pre-shipment in a predetermined frequency range A voltage is detected using at least one cell selected on the basis of intensity as a detection unit, a frequency component in a predetermined frequency range is extracted from the detected voltage, and the fuel cell stack is extracted from the size of the extracted frequency component Alternatively, a first control step of diagnosing the presence or absence of an abnormality in the power generation system including the fuel cell stack and reducing a current value taken out from the fuel cell stack until no signal indicating abnormality is output when the abnormality is detected And then performing a second control step for avoiding the abnormality based on the cause of the abnormality. Beam method of operation.

Images