Cell stack evaluation device and cell stack evaluation method

Abstract

To provide a cell stack evaluation device capable of easily achieving evaluation of cracks generated in a cell stack with high efficiency and precision.SOLUTION: A cell stack evaluation device includes an input unit and an estimation unit. The input unit receives state data related to a state of a cell stack. The estimation unit obtains crack position estimation data by estimating a position of a crack occurring in a plurality of electrochemical cells based on the state data input from the input unit. The estimation unit is configured to obtain the crack position estimation data using an evaluation model obtained by performing machine learning on a relation between the state data and a position of a crack occurring in the plurality of electrochemical cells.SELECTED DRAWING: Figure 1A

Description

[Technical Field]

[0001] An embodiment of the present invention relates to a cell stack evaluation device and a cell stack evaluation method. [Background technology]

[0002] The electrochemical device has an electrochemical cell configured such that an electrolyte membrane is sandwiched between a hydrogen electrode and an oxygen electrode. Among electrochemical cells, a solid oxide electrochemical cell using a solid oxide for the electrolyte membrane can be used as at least one of a solid oxide fuel cell (SOFC) and a solid oxide electrolysis cell (SOEC).

[0003] When a solid oxide electrochemical cell is used as an SOFC, for example, hydrogen supplied to the hydrogen electrode and oxygen supplied to the oxygen electrode react with each other through an electrolyte membrane under high-temperature conditions to generate electrical energy. On the other hand, when a solid oxide electrochemical cell is used as an SOEC, for example, water (water vapor) is electrolyzed under high-temperature conditions to generate hydrogen at the hydrogen electrode and oxygen at the oxygen electrode.

[0004] Generally, an electrochemical device includes a cell stack in which a plurality of electrochemical cells are stacked. The cell stack includes a plurality of separators and the like in addition to the plurality of electrochemical cells, and is configured so that each of the plurality of electrochemical cells is sandwiched between the plurality of separators in the stacking direction. In the cell stack, the plurality of electrochemical cells are electrically connected to each other via the separators and the like to increase the power generation output, etc. The cell stack is sandwiched between a pair of end plates, and the pair of end plates are fastened together using fastening members such as bolts, for example.

[0005] In the cell stack, the electrochemical cells are made of brittle materials such as ceramics and are thin. Therefore, the electrochemical cells may crack due to brittle fracture. If a crack occurs in the electrochemical cell, the hydrogen electrode gas flowing through the hydrogen electrode and the oxygen electrode gas flowing through the oxygen electrode mix in the electrochemical cell through the crack, resulting in a decrease in the performance of the electrochemical device. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6317710 [Non-patent literature]

[0007] [Non-Patent Document 1] NEDO Fuel Cell and Hydrogen Technology Development Roadmap 2022, [Retrieved April 23, 2024], Internet<https: / / www.nedo.go.jp / library / battery_hydrogen.html> [Non-patent document 2] Sato, Tomoyoshi et al., "Development of a transient electrochemical-mechanical coupled analysis system for solid oxide fuel cells using a general-purpose FEM software platform," Transaction of JSCES 2017, [Retrieved April 23, 2024], Internet <URL:https: / / www.jstage.jst.go.jp / article / jsces / 2017 / 0 / 2017_20170004 / _article / -char / ja / > Summary of the Invention [Problem to be solved by the invention]

[0008] When designing electrochemical devices, evaluation of cracks occurring in cell stacks is performed. Evaluation of cracks in cell stacks is performed, for example, by a mop-up test. However, performing a mop-up test requires a lot of time and costs a lot of money. For this reason, there is a demand to reduce the frequency of mop-up tests.

[0009] Numerical analysis using the finite element method (FEM) or homogenization method is one way to reduce the frequency of mop-up tests. However, to perform highly accurate evaluations using the FEM, the cell stack to be evaluated must be divided into small elements with simple structures, which increases the computational processing time required for the analysis. Specifically, it takes time to calculate the stress balance at the contact points between components with different rigidities (e.g., separators and current collectors) in the cell stack. While analysis using the homogenization method is expected to reduce computational processing time, it is difficult to accurately determine the average material constants for each component of the cell stack, making it difficult to accurately evaluate cell stack cracking. In particular, the above-mentioned problems become more apparent as the number of electrochemical cells stacked increases due to demands for larger capacity.

[0010] Due to the above circumstances, it has not been easy to efficiently and accurately evaluate cracks that occur in a cell stack when designing an electrochemical device.

[0011] Therefore, the problem that the present invention aims to solve is to provide a cell stack evaluation device and a cell stack evaluation method that can easily realize efficient and highly accurate evaluation of cracks that occur in a cell stack. [Means for solving the problem]

[0012] A cell stack evaluation device according to an embodiment is configured to evaluate a cell stack in which a plurality of electrochemical cells, each having an electrolyte membrane interposed between a hydrogen electrode and an oxygen electrode, are sandwiched between a plurality of separators in the stacking direction, and hydrogen electrode gas flows through the hydrogen electrode while oxygen electrode gas flows through the oxygen electrode. The cell stack evaluation device according to an embodiment includes an input unit and an estimation unit. The input unit receives state data relating to the state of the cell stack. The estimation unit obtains crack position estimation data by estimating the positions of cracks that have occurred in the plurality of electrochemical cells based on the state data input from the input unit. The estimation unit is configured to obtain the crack position estimation data using an evaluation model obtained by machine learning about the relationship between the state data and the positions of cracks that have occurred in the plurality of electrochemical cells. [Brief explanation of the drawings]

[0013] [Figure 1A] FIG. 1A is a diagram schematically showing a main part of an electrochemical device 1 to be evaluated in a cell stack evaluation device according to an embodiment (Y1-Y1 portion in FIG. 1B). [Figure 1B] FIG. 1B is a diagram schematically showing a main part of an electrochemical device 1 to be evaluated in the cell stack evaluation device of the embodiment (Z1-Z1 part in FIG. 1A). [Figure 2] FIG. 2 is a functional block diagram that schematically shows a cell stack evaluation device 80 according to the embodiment. [Figure 3] FIG. 3 is a flow diagram showing the operation of the cell stack evaluation device 80 according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] [A] Configuration of electrochemical device 1 Before describing the cell stack evaluation device of the embodiment, an electrochemical device 1 that is the subject of evaluation by the cell stack evaluation device will be described first.

[0015] 1A and 1B are diagrams that schematically show the main parts of an electrochemical device 1 that is the subject of evaluation in a cell stack evaluation device according to an embodiment.

[0016] 1A, the longitudinal direction is the vertical direction z, the horizontal direction is the first horizontal direction x, and the direction perpendicular to the paper surface is the second horizontal direction y that is perpendicular to the vertical direction z and the first horizontal direction x. Fig. 1A is a side cross-sectional view of the electrochemical device 1, showing a portion corresponding to the plane (xz plane) of the Y1-Y1 portion in Fig. 1B.

[0017] 1B, the longitudinal direction is the second horizontal direction y, the lateral direction is the first horizontal direction x, and the direction perpendicular to the paper surface is the vertical direction z. Fig. 1B is a top view of the electrochemical device 1, showing a portion corresponding to the plane (xy plane) of the Z1-Z1 portion in Fig. 1A.

[0018] As shown in FIGS. 1A and 1B, the electrochemical device 1 includes an electrochemical cell 10, current collectors 21 and 22, and separators 31 and 32, and is configured to perform at least one of power generation and electrolysis.

[0019] Although not shown, the electrochemical device 1 includes a cell stack having a plurality of electrochemical cells 10, each sandwiched between a plurality of separators (reference numerals 31 and 32 in FIG. 1A ) in the stacking direction (vertical direction z in FIG. 1A ). In the cell stack, current collectors (reference numerals 21 and 22 in FIG. 1A ) are interposed between each of the electrochemical cells 10 and each of the separators (reference numerals 31 and 32 in FIG. 1A ). In the cell stack, the electrochemical cells 10 are electrically connected in series to increase power generation output, etc. The cell stack is sandwiched between a pair of end plates (not shown) in the vertical direction z, and the pair of end plates is fastened using fastening members such as bolts. The cell stack is also electrically connected to a pair of bus bars (not shown). A current is supplied to the cell stack via the pair of bus bars during electrolysis, and a current is extracted via the pair of bus bars during power generation.

[0020] Each component of the electrochemical device 1 will be described below in order.

[0021] [A-1] Electrochemical cell 10 The electrochemical cell 10 is, for example, a rectangular flat plate type, and includes an electrolyte membrane 110, a hydrogen electrode 111, and an oxygen electrode 112, with the electrolyte membrane 110 interposed between the hydrogen electrode 111 and the oxygen electrode 112. Here, the electrochemical cell 10 is, for example, a hydrogen electrode-supported type (fuel electrode-supported type), with the electrolyte membrane 110 and the oxygen electrode 112 sequentially stacked on top of the hydrogen electrode 111, which functions as a support.

[0022] In the electrochemical cell 10, the electrolyte membrane 110 is a 2- The electrolyte membrane 110 is made of an ion-conductive solid oxide (e.g., yttria-stabilized zirconia (YSZ)) that is permeable to oxygen. The electrolyte membrane 110 is configured to be denser than the hydrogen electrode 111 and the oxygen electrode 112. The hydrogen electrode 111 is made of a porous electrical conductor (e.g., a cermet formed using nickel particles and ceramic particles such as YSZ). The oxygen electrode 112 is made of a porous electrical conductor (e.g., a perovskite oxide such as LaSrMnO3).

[0023] As described above, in the electrochemical cell 10, the electrolyte membrane 110, the hydrogen electrode 111, and the oxygen electrode 112 are each made of a brittle material such as a ceramic material.

[0024] [A-2] Current collector 21, 22 The current collector 21 is provided on the underside of the hydrogen electrode 111 that constitutes the electrochemical cell 10. The current collector 21 has a mesh or porous structure and is configured to allow permeation of the hydrogen electrode gas consumed or generated at the hydrogen electrode 111. The current collector 21 is also made of a metal material such as nickel, and electrically connects the hydrogen electrode 111 to the separator 31 located below the hydrogen electrode 111.

[0025] The current collector 22 is provided on the upper surface of the oxygen electrode 112 that constitutes the electrochemical cell 10. Like the current collector 21, the current collector 22 has a mesh structure or a porous structure, and is configured to allow the oxygen electrode gas consumed or generated in the oxygen electrode 112 to pass through. The current collector 22 is also made of a metal material such as silver, and electrically connects the oxygen electrode 112 and the separator 32 located above the oxygen electrode 112.

[0026] [A-3] Separators 31, 32 The separator 31 is made of a conductive material such as metal and is disposed on the hydrogen electrode 111 side of the electrochemical cell 10. The separator 31 has an accommodation space K31. The accommodation space K31 is formed in a central portion of the upper surface of the separator 31. The accommodation space K31 has a planar shape, for example, a rectangular recess, and is configured to accommodate the current collector 21 and the electrochemical cell 10. In this example, the accommodation space K31 accommodates the electrolyte membrane 110 and the hydrogen electrode 111 of the electrochemical cell 10. The separator 31 is also provided with a hydrogen electrode gas supply port F311, a hydrogen electrode gas flow path F31, and a hydrogen electrode gas discharge port F312, and is configured so that hydrogen electrode gas supplied from the hydrogen electrode gas supply port F311 passes through the hydrogen electrode gas flow path F31 and is then discharged from the hydrogen electrode gas discharge port F312. Here, the hydrogen electrode gas flow path F31 is, for example, a linear groove, and is formed on the support surface (bottom surface) that supports the hydrogen electrode 111 in the accommodation space K31.

[0027] Like the separator 31, the separator 32 is made of a conductive material such as metal and is disposed on the oxygen electrode 112 side of the electrochemical cell 10. The separator 32 is also provided with an oxygen electrode gas supply port F321, an oxygen electrode gas flow path F32, and an oxygen electrode gas outlet F322, and is configured so that oxygen electrode gas supplied from the oxygen electrode gas supply port F321 passes through the oxygen electrode gas flow path F32 and is then discharged from the oxygen electrode gas outlet F322. The oxygen electrode gas flow path F32 is provided on the lower surface of the separator 32 facing the upper surface of the oxygen electrode 112. The oxygen electrode gas flow path F32 is, for example, a linear groove formed perpendicular to the linear groove constituting the hydrogen electrode gas flow path F31.

[0028] The separators 31 and 32 are formed of a material having higher rigidity than the current collectors 21 and 22 .

[0029] [A-5] Sealant 40 A sealing material 40 is interposed between the separator 31 and the separator 32. The sealing material 40 is a frame-shaped plate body with an opening K40 formed in the center, and the oxygen electrode 112 and the current collector 22 are housed inside the opening K40. The sealing material 40 also includes a portion that protrudes inward above the housing space K31, and this protruding portion covers the gap and contacts the upper surface of the electrolyte membrane 110.

[0030] The seal material 40 is configured to seal the space between the separator 31 and the separator 32. The seal material 40 is also configured to electrically insulate the space between the separator 31 and the separator 32.

[0031] Here, the sealing material 40 includes a sealing material main body 400 and sealing material metal plate portions 411 and 412. In the sealing material 40, the sealing material main body 400 is a gasket made of an insulating material such as mica or vermiculite.

[0032] The sealing metal plate portions 411 and 412 are made of plates of a metal material, and are arranged so as to sandwich the sealing material main body 400 in the stacking direction.

[0033] In the seal 40, the seal metal plate portion 411 and the seal metal plate portion 412 are made of a material having higher rigidity than the seal main body 400.

[0034] [B] Configuration of cell stack evaluation device 80 FIG. 2 is a functional block diagram that schematically shows a cell stack evaluation device 80 according to the embodiment.

[0035] As shown in FIG. 2, the cell stack evaluation device 80 of this embodiment has an input unit 81, an estimation unit 83, and an output unit 85, and is configured to evaluate a cell stack in which multiple electrochemical cells 10 are stacked in the above-mentioned electrochemical device 1.

[0036] [B-1] Input section 81 The input unit 81 includes, for example, an input interface, and receives state data D81 (state variables) relating to the state of the cell stack.

[0037] [B-2] Estimation part 83 The estimation unit 83 is configured by, for example, a computer, and executes a program to perform an estimation process to estimate the positions of cracks that occur in the plurality of electrochemical cells 10 that constitute the cell stack in the electrochemical device 1. Based on the state data D81 input from the input unit 81, the estimation unit 83 executes an estimation process to estimate the positions of cracks that occur in the plurality of electrochemical cells 10, thereby obtaining crack position estimation data D83.

[0038] Here, the estimation unit 83 has an evaluation model 830 stored in a storage device, and uses the evaluation model 830 to determine crack position estimation data D83. That is, the estimation unit 83 estimates the position of a crack by applying the state data D81 input from the input unit 81 to the evaluation model 830, and determines the crack position estimation data D83. In the estimation unit 83, the evaluation model 830 is generated in advance by machine learning using the relationship between the state data D81 and the positions of cracks that occur in the multiple electrochemical cells 10 as learning data (teaching data).

[0039] [B-3] Output section 85 The output unit 85 is, for example, an information output device including a display, and is configured to output information relating to the crack position estimation data D83 acquired by the estimation unit 83 onto the screen of the display.

[0040] In addition to the display, the output unit 85 may further include an information output device such as a printer for printing and displaying output.

[0041] [C] Operation of the cell stack evaluation device 80 The operation of the cell stack evaluation device 80 (see FIG. 2) of this embodiment will be described.

[0042] FIG. 3 is a flow diagram showing the operation of the cell stack evaluation device 80 according to the embodiment.

[0043] As shown in FIG. 3, in the cell stack evaluation device 80 of this embodiment (see FIG. 2), an input step (ST81), an estimation step (ST83), and an output step (ST85) are executed in sequence.

[0044] Each step will be explained using FIG. 2 together with FIG.

[0045] [C-1] Input step (ST81) The input step (ST81) is executed by the input unit 81 (see FIG. 2) of the cell stack evaluation device 80. In the input step (ST81), state data D81 (state variables) relating to the state of the cell stack are input to the input unit 81. For example, the state data D81 is input by an operator's operation.

[0046] In this embodiment, load data D811, temperature data D812, and gas component data D813 are input as the state data D81.

[0047] In the state data D81, load data D811 is data relating to the load applied to the cell stack, and includes press pressure data D811a, gas pressure data D811b, and gas pressure difference data D811c.

[0048] Specifically, the press pressure data D811a is data relating to the press pressure (P1) applied to the cell stack in the stacking direction.

[0049] The gas pressure data D811b is data relating to the hydrogen electrode gas pressure (PG1) of the hydrogen electrode gas supplied to the hydrogen electrode 111 and the oxygen electrode gas pressure (PG2) of the oxygen electrode gas supplied to the oxygen electrode 112 in the cell stack.

[0050] The gas pressure difference data D811c is data relating to the pressure difference (ΔPG=PG1-PG2) between the hydrogen electrode gas pressure (PG1) of the hydrogen electrode gas supplied to the hydrogen electrode 111 and the oxygen electrode gas pressure (PG2) of the oxygen electrode gas supplied to the oxygen electrode 112 in the cell stack.

[0051] In the state data D81, the temperature data D812 is data relating to the operating temperature (T) of the cell stack.

[0052] In the state data D81, the gas component data D813 is data relating to the components of the hydrogen electrode gas at the hydrogen electrode 111 and the components of the oxygen electrode gas at the oxygen electrode 112 in the electrochemical cell 10 of the cell stack. When the electrochemical cell 10 is used as a fuel cell (SOFC), the gas component data D813 is data relating to the partial pressure (P H2 ), and the partial pressure of oxygen in the oxygen electrode gas supplied to the oxygen electrode 112 (P O2 On the other hand, when the electrochemical cell 10 is used as an electrolysis cell (SOEC), the partial pressure of hydrogen in the hydrogen electrode gas containing hydrogen generated at the hydrogen electrode 111 (P H2 ), and the partial pressure of oxygen in the oxygen electrode gas containing oxygen generated at the oxygen electrode 112 (P O2 ) data.

[0053] [C-2] Estimation step (ST83) The estimation step (ST83) is executed by an estimation unit 83 (see FIG. 2) of the cell stack evaluation device 80. In the estimation step (ST83), the estimation unit 83 executes an estimation process to estimate the positions of cracks that occur in the multiple electrochemical cells 10 based on the state data D81 input to the input unit 81 in the input step (ST81), and crack position estimation data D83 is obtained.

[0054] The crack position estimation data D83 is obtained by the estimation unit 83 using the evaluation model 830. That is, the estimation unit 83 estimates the position of the crack by applying the state data D81 input from the input unit 81 to the evaluation model 830, and obtains the crack position estimation data D83.

[0055] As described above, the evaluation model 830 is generated in advance by machine learning using the relationship between the state data D81 and the positions of cracks that occur in the multiple electrochemical cells 10 as learning data (teaching data). The learning data is a data set in which the state data D81, including load data D811, temperature data D812, and gas component data D813, is associated with data related to the positions of cracks. The learning data may be, for example, data obtained by analysis using a simulator, or may be data obtained by a mop-up test.

[0056] The evaluation model 830 is configured, for example, by a neural network consisting of an input layer, an intermediate layer (hidden layer), and an output layer (not shown). In the neural network configuring the evaluation model 830, the values ​​of the load data D811, temperature data D812, and gas component data D813 input as state data D81 are input to each node configuring the input layer, and then output from each node configuring the input layer to each node configuring the intermediate layer. In each node configuring the intermediate layer, a weighted addition is performed on the input values. Then, the value determined in each node configuring the intermediate layer is output to each node configuring the output layer. In the evaluation model 830, the values ​​of the weighting coefficients used in the weighted addition are adjusted by machine learning.

[0057] The value obtained at each node constituting the output layer in the neural network that constitutes the evaluation model 830 corresponds to, for example, the probability that a crack will occur at each location of the multiple electrochemical cells 10 that constitute the cell stack. For example, when the value obtained at a node constituting the output layer in the neural network exceeds a predetermined threshold, the estimation unit 83 estimates that a crack will occur at the location corresponding to that node.

[0058] [C-3] Output step (ST85) The output step (ST85) is executed by the output unit 85 (see FIG. 2) of the cell stack evaluation device 80. In the output step (ST85), information related to the crack position estimation data D83 acquired in the estimation step (ST83) is output, for example, to a display screen. Specifically, information on the locations where cracks occur in the multiple electrochemical cells 10 that make up the cell stack is displayed on the display screen based on the crack position estimation data D83.

[0059] [D] Summary As described above, in this embodiment, status data D81 relating to the status of the cell stack is input to the input unit 81. The locations of cracks occurring in the plurality of electrochemical cells 10 are estimated by the estimation unit 83 based on the status data D81 input from the input unit 81, and crack position estimation data D83 is obtained. The crack position estimation data D83 is obtained using an evaluation model 830 obtained by machine learning on the relationship between the status data D81 and the locations of cracks occurring in the plurality of electrochemical cells. Information relating to the crack position estimation data D83 is output by the output unit 85.

[0060] Therefore, in this embodiment, it is possible to easily perform efficient and highly accurate evaluation of cracks in the cell stack when designing an electrochemical device.

[0061] In this embodiment, the state data D81 includes load data D811, temperature data D812, and gas component data D813. Here, the load data D811 includes press pressure data D811a, gas pressure data D811b, and gas pressure difference data D811c. Each piece of data significantly affects the location of cracks that occur in multiple electrochemical cells. Therefore, it is possible to estimate the location of cracks in the cell stack with high accuracy.

[0062] <Other> Although several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be embodied in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included within the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents as defined in the claims. [Explanation of symbols]

[0063] 1: electrochemical device, 10: electrochemical cell, 21: current collector, 22: current collector, 31: separator, 32: separator, 40: sealing material, 80: cell stack evaluation device, 81: input unit, 83: estimation unit, 85: output unit, 110: electrolyte membrane, 111: hydrogen electrode, 112: oxygen electrode, 400: sealing material main body, 411: sealing material metal plate part, 412: sealing material metal plate part, 830: evaluation model, D81: state data, D811: load Weight data, D811a: Press pressure data, D811b: Gas pressure data, D811c: Gas pressure difference data, D812: Temperature data, D813: Gas component data, D83: Crack position estimation data, F31: Hydrogen electrode gas flow path, F311: Hydrogen electrode gas supply port, F312: Hydrogen electrode gas outlet, F32: Oxygen electrode gas flow path, F321: Oxygen electrode gas supply port, F322: Oxygen electrode gas outlet, K31: Storage space, K40: Opening

Claims

1. A cell stack evaluation device for evaluating a cell stack configured such that a plurality of electrochemical cells, each having an electrolyte membrane interposed between a hydrogen electrode and an oxygen electrode, are sandwiched between a plurality of separators in the stacking direction, and hydrogen electrode gas flows through the hydrogen electrode and oxygen electrode gas flows through the oxygen electrode, an input unit to which status data relating to the status of the cell stack is input; an estimation unit that estimates positions of cracks occurring in the plurality of electrochemical cells based on the state data input from the input unit, thereby obtaining crack position estimation data; and The estimation unit is configured to obtain the crack position estimation data using an evaluation model obtained by machine learning about the relationship between the state data and positions of cracks occurring in the plurality of electrochemical cells. Cell stack evaluation device.

2. The status data is Load data relating to a load applied to the cell stack; and temperature data relating to the operating temperature of the cell stack; gas component data relating to the components of the hydrogen electrode gas at the hydrogen electrode and the components of the oxygen electrode gas at the oxygen electrode; At least including The cell stack evaluation device according to claim 1 .

3. The load data is press pressure data relating to a press pressure applied to the cell stack in the stacking direction; gas pressure data relating to the hydrogen electrode gas pressure of the hydrogen electrode gas supplied to the hydrogen electrode and the oxygen electrode gas pressure of the oxygen electrode gas supplied to the oxygen electrode in the cell stack; gas pressure difference data relating to the pressure difference between the hydrogen electrode gas pressure of the hydrogen electrode gas supplied to the hydrogen electrode and the oxygen electrode gas pressure of the oxygen electrode gas supplied to the oxygen electrode in the cell stack; Including, The cell stack evaluation device according to claim 2 .

4. An output unit 85 that outputs information about the crack position estimation data acquired by the estimation unit Including, The cell stack evaluation device according to claim 1 .

5. A cell stack evaluation method for evaluating a cell stack in which a plurality of electrochemical cells, each having an electrolyte membrane interposed between a hydrogen electrode and an oxygen electrode, are sandwiched between a plurality of separators in the stacking direction, and hydrogen electrode gas flows through the hydrogen electrode and oxygen electrode gas flows through the oxygen electrode, an input step in which status data relating to the status of the cell stack is input; an estimation step of estimating positions of cracks occurring in the plurality of electrochemical cells based on the state data input in the input step, thereby obtaining crack position estimation data; and the estimation step obtains the crack position estimation data using an evaluation model obtained by machine learning about the relationship between the state data and positions of cracks occurring in the plurality of electrochemical cells; The status data is Load data relating to a load applied to the cell stack; and temperature data relating to the operating temperature of the cell stack; gas component data relating to the components of the hydrogen electrode gas at the hydrogen electrode and the components of the oxygen electrode gas at the oxygen electrode; At least including Cell stack evaluation method.

Images